Every morning, Camille opens his terminal. Eight new pull requests are waiting for him. Created, reviewed, and pre-approved while he was asleep.

Not by a team. By his agents.

In this episode of Build with AI, Tanguy Goretti (CTO at Hexa) sits down with Camille Epitalon, CTO and co-founder of Verso, a startup he built entirely solo on the technical side in three months, with over twenty paying enterprise clients already on board. Camille opens his terminal live and walks through, brick by brick, the setup that makes it all possible.

Verso: A full-AI qualitative research platform

Before diving into the setup, a word on what Verso actually does, because it’s inseparable from how Camille codes.

Verso is an AI-native qualitative research platform. In practice: where agencies like Kantar or Ipsos used to charge €100,000 and three months of lead time to conduct qualitative consumer interviews, Verso delivers the same output in three days for a few thousand euros.

The product covers the entire chain: an AI agent co-designs the interview guide with the client, other agents recruit participants, an AI moderator conducts the interviews (video, audio, screen share, emotion analysis), and researcher agents produce a dynamic report from the raw data. The client can then query that data through a built-in chat interface.

The philosophy: giving codebase ownership to your agents

From the very first minutes, Camille lays out the mental framework for everything that follows:

“From day one, I told myself: I’m handing ownership of my codebase to my agents. The codebase belongs to my agents.”

It sounds radical. But Camille is precise about what it means and more importantly, what it doesn’t mean.

Delegating ownership doesn’t mean letting go of the wheel. It means investing heavily in what he calls harness engineering: everything that surrounds the codebase so that agents operate within a safe, consistent, and opinionated framework.

Two conditions are non-negotiable, according to him:

1. Solid foundations in the code. Clear abstract classes, well-defined contracts, enforced patterns. For example, all his repos inherit from an OrgScopedRepo: which ensures that when an agent creates a new repo, it automatically picks up all the security and authentication abstractions. The agent doesn’t need to be told the rules: they’re baked into the structure of the code.

2. Well-equipped agents. Not just with instructions, but with the right tools to observe the environment: logs, tracking, GitHub, CI. An autonomous agent needs to be able to connect to everything it needs to make decisions without constantly asking its human.

The live setup: 6 worktrees, aliases, a terminal as cockpit

Camille moved from VS Code and Cursor to... his terminal.

His setup is built around 6 permanent worktrees, each associated with a color in his terminal (via iTerm). When he opens a profile, the worktree automatically re-syncs with main. The entire lifecycle (reset, sync, opening a PR) is managed through aliases. To launch Claude Code, he types c. Codex: x. Full dev environment: d.

The metaphor he uses is vivid: “My six worktrees are like six different engineers. I try to avoid having them work on the same thing.”

In practice, each worktree maps to one PR. Camille works async, he kicks off a task, switches to another worktree while the first one runs, and comes back to check the result later. Six instances in parallel, but never all active at the same time.

The cron jobs that code through the night

This is the part that raises eyebrows.

Every two hours, and especially overnight, cron jobs launch Claude or Codex agents in headless mode from Camille’s Mac. These agents follow predefined skills, do their work, and open pull requests on GitHub. In the morning, Camille just has to review them.

These cron jobs are primarily focused on clean code and refactoring. A few concrete examples:

• Backend audit: reads the Python/FastAPI guidelines, produces a report prioritized by criticality and fix complexity, then picks the best value/effort trade-off to create a targeted PR.

• Frontend audit: same principle on the React/TypeScript side.

• Skills sync: checks every day that the .claude and .codex config folders are in sync - if a skill was updated on one side but not the other, a PR is automatically opened.

• Simplify skill: a Claude built-in skill that identifies over-complex areas in the codebase and proposes simplifications.

The logic: rather than blocking a large refactoring project, agents chip away every night at small, targeted improvements. Day after day, the codebase cleans itself without friction.

“Verso is the codebase that never sleeps.”

The Ralph Loop: launch a feature before bed

For heavier features, Camille uses the Ralph Loop.

The concept: a bash loop running continuously in the terminal. At each iteration, it reads a Markdown spec written upfront, consults a tracking file that records where the implementation stands, spins up a new agent (Claude Code or Codex) to handle only the next step, and updates the tracking file with learnings and results.

Each agent operates with a fresh context and a limited scope, avoiding context bloat and drift over long tasks. The source of truth is never in the context window: it’s in the filesystem.

In practice: Camille writes the spec with an agent’s help before heading to lunch or going to bed. He launches the ralph loop. When he comes back, the feature is implemented, or close to it.

Camille notes that some sessions can burn through all his API credits in a single night on ambitious tasks.

The automated review chain

No PR ships without going through multiple layers of review. Here’s the full chain:

1. Auto-review during implementation. The Implement skill automatically calls Review Current Work at each step. The agent self-reviews before opening a PR.

2. Cross-review between agents. Claude Code and Codex review each other’s work. Camille doesn’t necessarily read the diff himself - he asks Codex to review Claude’s work and vice versa, via the Check Reviews skill, which triages comments and proposes fixes.

3. External AI reviewers. Three tools run on every PR: BugBot, Cubic (a YC startup that Camille says is taking off fast on benchmarks), and Code Rabbit (which he finds less impressive today but keeps out of habit). After each comment, a loop mechanism checks the PR 30 minutes after the push, triages the feedback, fixes what can be fixed, and pushes again.

4. Final human review. Camille steps in last, for high-impact decisions or complex trade-offs. The skills are explicit about this: “Don’t make one-way-door decisions without me.”

The cost of all these tools? A few hundred euros a month. Camille is unequivocal: it’s a non-negotiable expense.

Live Browser QA

To validate front-end changes, Camille uses a browser QA skill inspired by Vercel’s. The agent connects to the local app (each worktree runs its own instance with its own DB), navigates to the given URL, takes screenshots, clicks, interacts, and produces a report detailing what works and what doesn’t.

In the live demo, you can watch the agent open Chrome, auto-login to Verso, find the right component, test it, and validate that a video playback speed control button works correctly.

The current limitation: screenshots don’t capture animations. Camille thinks the real revolution in agentic QA will come when models can read video in real time.

Business skills: the whole team is AI-first

An often-overlooked point: at Verso, agents aren’t reserved for developers.

From day one of onboarding, everyone, including the business team, has access to Claude Code, Codex, and the terminal. Business-specific skills have been built for them: drafting an email, preparing a client’s context in the knowledge base, writing a LinkedIn post, or aggregating all call transcripts and emails from a client to pre-fill study templates.

There’s also the Teach skill: when Camille just accomplished something manually with an agent, he can immediately ask it to document what it just learned as a reusable skill. It’s the self-improvement mechanism of the harness itself.

The agnosticism principle: never get locked in

Camille emphasizes a structural point: his setup has to work regardless of the provider.

Skills are nearly identical between Claude Code and Codex (with minor format adaptations), and a cron job checks every day that both config folders are in sync. When Anthropic has outages he switches to Codex without friction.

“Tomorrow, if another provider is better, I’ll switch with pleasure. Being agnostic and not locked in is pretty important.”

Takeaways for ×10 (or ×100) developer productivity

In closing, Camille and Tanguy summarize the principles they apply at Hexa and Verso:

Invest in foundations first. Agents build on what they see. If the code structure is ambiguous or inconsistent, AI will drift in the wrong direction and it compounds. Clear patterns, well-thought-out abstractions, guidelines documented as Markdown files in the codebase: that’s what determines the quality of everything produced afterward.

Run regular audits. Periodically, ask the agent to cold-analyze the codebase: what are the main separation-of-concerns violations? Where is there code duplication? Which areas have drifted from the guidelines? It’s an effective way to identify technical debt before it accumulates.

Engineer the harness hard. Skills, hooks, agents, cron jobs: everything surrounding the codebase deserves as much attention as the codebase itself. It’s an investment with exponential returns.

Don’t panic about new tools. A new “game-changer” tool drops every week on LinkedIn. Tools change. Principles don’t.

“Vibe coding is the worst name for this practice. It’s real engineering work, defined workflows, clear foundations. It’s not ‘go with the vibe’ at all.” - Camille Epitalon

The bottom line

What Camille built at Verso isn’t just an efficient dev setup. It’s a way of working where agents are treated as real collaborators - with clear responsibilities, a well-defined framework, and progressive autonomy.

The result: an enterprise product in production in three months, built alone. Paying clients. And nights where the code keeps improving while he sleeps.

Most founders don’t think much about brand in the early days. It doesn’t feel urgent - building, shipping, and fundraising take priority. Brand can seem like something for later, once everything else is in place.

But you already have a brand, whether you’ve thought about it or not.

Customers, investors, partners, they’re forming impressions based on what they see: your landing page, your first deck, a job post, even your LinkedIn bio. Whether or not there was intention behind those things, people are making snap judgments. Who is this company? Do they seem credible? Do I like them? What does using their product say about me?

That’s brand: the feeling people get from what you put into the world. And
in the early days, when attention is scarce and trust is fragile, those signals really matter. If they’re generic or inconsistent, people move on. But if there’s a thread - something clear, confident, emotionally resonant - people are more likely to take you seriously, and even start believing in what you’re building.

There was a time when being modern and simple was enough. Not anymore. Everyone can look modern now. And with AI and faster dev tools, any existing B2B SaaS can be copied or rebuilt in months. As product quality evens out, it’ll be your brand - and the people behind it - that makes people choose you over a competitor.

In this chapter, we’ll show you how to build a strong early-stage brand. From defining your identity and voice to bringing it to life through design, messaging, founder presence, and strategic communication across the right channels.

Why brand matters in the early stage

Brand isn’t just a late-stage concern or a luxury for when you have a marketing team. Your landing page, your first pitch deck, a job post, even your LinkedIn bio - it’s all brand, and they leave important first impressions. You’re trying to earn attention, attract top candidates, get investors excited, convince early customers to take a bet on you… all of that hinges on what they feel when they encounter your company.

There are two primary ways you shape that feeling:

1. Visual identity: how you look. It’s your product interface, your landing page, your logo, your color palette.

2.Communications: how you sound. Across your website copy, social media, investor emails, and even the way you talk to users.

It’s easy to dismiss the importance of brand at the start, especially in B2B. But in B2B, professionals are still human. They don’t show up at work, switch off their emotional brain, and put on a purely rational hat. Professional purchases are human purchases, influenced by feelings: What does this choice say about me? How does it make me feel? What will it signal to my team or industry? When multiple products could meet their needs, the brand that makes them feel something will win.

When the functional differences between products are slim, it’s the brand that evokes the right emotion that wins.

That’s why at Hexa, we push founders to think about brand early. But it’s not about creating a polished, expensive deliverable or 50 pages of brand guidelines no one reads. It should be a focused, intentional effort to signal who you are and why someone should care. At this stage, you’re building a “just good enough” brand. One that creates an emotional hook and helps you stand out.

Defining your brand

Get clear on who you are - really

Consistency is the foundation of every strong brand, but you can’t be consistent if you don’t know exactly what you’re being consistent about.

This is a challenge most early-stage founders face: things move fast, priorities shift, and your story can evolve week to week. One day you’re focused on one angle, the next day it’s something entirely different, and that’s perfectly normal.

But from a branding perspective, it’s essential to commit to a single, clear narrative, at least for a given moment in time. If you try to say everything at once, you risk saying nothing at all. Consistency is what makes you memorable - it helps your audience associate your brand with one strong idea, rather than getting lost in a swirl of mixed messages.

The solution is simple: document your core brand elements in one place to force clarity on a single version of who you are. You can refer back to this document to keep every company action aligned with the bigger picture, from writing your landing page copy to making strategic decisions.

Here’s a template to fill out:

“I think our secret weapon early on was internal alignment. We didn’t intentionally build a brand back then, but looking back, it was strong from the start. We cared deeply about what we were building, and that came through.

I remember sharing a deck early on with a simple diagram, three concentric circles: the team, our customers, and the broader community. The idea was to show how our values, like remote work, CO2 impact, and work-life balance, could radiate outward. It was a small gesture, but it resonated and helped shape how we showed up.

Over time, we heard again and again that customers chose Slite because of who we were and what we stood for, not just the product.”

– Christophe Pasquier, CEO & Cofounder of Slite

Find your brand feeling

Your core brand elements (above) are crucial for alignment and consistency, but your audience will never read your vision statement. Instead, they’ll experience your brand as a feeling. Simply put, your brand feeling is the emotional response you want people to have when they interact with your company. It’s not what your product does, it’s how it makes people feel.

As an early-stage founder, defining this feeling early gives you a major advantage. While you may not have the marketing budget to compete with established players, you can create a distinctive emotional connection that helps you stand out from day one - whether through your product, website, copy, or anything else you put into the world.

You probably won’t get this perfect on the first try, and that’s okay. For early-stage startups, your brand feeling doesn’t need to be polished or permanent. It’s just a starting point that gives you enough direction to be consistent while you’re still figuring things out. You’ll refine it as you learn more about your customers.

Examples:

Spendesk

Spendesk tapped into a feeling of liberation – they free finance leaders from being perceived as mere “spreadsheet people” and elevate them to strategic business partners. Users buy a vision of themselves as valued, strategic leaders in their organizations, not just spend management software.

“We had a few strong convictions early on that we held onto firmly and that ended up shaping our brand in meaningful ways. One clear example was around color: I remember working with an agency on our logo and overall visual identity, and at the time, most companies in our space were using very similar, fairly conservative color palettes.

We decided to break away from that by choosing purple. A choice that felt bold and different back then, even though today purple has beco- me much more common in tech branding. The idea was to make our Spendesk cards stand out: we didn’t want them to look like the usual, traditional black corporate cards that conveyed a very formal image; instead, we wanted something more vibrant and joyful. It wasn’t necessarily the “safe” or conventional choices but those bets paid off, helping us build a brand identity that felt memorable.

At some point, it became more than just branding: people would lite- rally pull out their purple Spendesk cards and proudly say, “I’m with Spendesk,” which for us was a clear sign that the brand had truly reso- nated.”

– Rodolphe Ardant, cofounder & CEO at Spendesk

Swan

Swan evokes a feeling of quiet empowerment – they help businesses offer financial services without surrendering their own brand identity to a bank. Their approach is built on the philosophy that “it’s about you, not us.” While traditional financial services dominate the experience, Swan steps into the background, allowing businesses to remain authentically themselves while gaining new capabilities. They’ve transformed what could be a purely technical service into something more meaningful: the ability to extend your brand into financial experiences without compromise.

“It’s easy to focus solely on things like product in the early days, but investing in brand early can be just as important. At Swan, one of the most impactful decisions we made was hiring a key brand leader early on. That hire helped shape Swan’s tone of voice and identity in a way that made the brand stand out, especially in a fintech space where most companies sound the same.”

– Nicolas Benady, cofounder & CEO at Swan

Turning it into a voice

Once you’ve defined the feeling you want to create, choose 2–3 adjectives that describe how your brand should speak to evoke it. For example, if your brand feeling is creative freedom (like Figma), your voice might sound: encouraging, approachable and a bit playful.

Each adjective should support the emotional experience you’re aiming for. “Encouraging” language helps people feel more free and capable. “Playful” gives people permission to loosen up a bit. Together, they reinforce the feeling you’re trying to create.

You don’t need a full set of brand guidelines at this stage, just a few clear voice adjectives can provide enough direction to shape your tone across everything: landing page copy, social posts, job descriptions, even how you respond to customer support tickets.

Bringing your brand to life - Visual Identity

One of the biggest worries founders have is: does building a good brand mean spending a ton of money? At this stage, no, it really doesn’t.

You can DIY 90-100% of your early brand, which makes sense when things are still evolving. Being scrappy now doesn’t mean settling for a sloppy result, it just means being smart about what actually matters.

Your name

While not visual in itself, your name is a crucial element that will strongly influence your visual identity - particularly the logo, which in many cases needs to visually represent the name. Naming your company can easily send you into endless loops. Yes, it needs to tick a few boxes, but not as many as you might think. Founders often put way too much pressure on finding the “perfect” name. The truth is that unless it’s actively bad, people get used to it. And if you do the brand work right, you’ll create meaning around it over time. The goal is finding a ‘good enough’ name that you can build on.

Here’s a quick breakdown of what your name does and doesn’t need to be:

“The name Swan came to us quickly and it just felt right. It captured the essence of what we wanted to build, something elegant but powerful. It also fit the naming trends we admired at the time: short, simple, starting with ‘S’ like Swile, Stripe, Shine. But more than that, it had emotional resonance. It was a name people would remember and feel something about.”

– Nicolas Benady, cofounder & CEO at Swan

Where and how to show up

Early on, your visual identity doesn’t need to be fancy, just clear and consistent. A simple logo and a basic color palette will take you a long way.

You can absolutely do this yourself using tools like Figma or Canva. Choose 1–2 signature colors that stand out in your space, pick a clean font, and keep it consistent across everything. If you’re using a website template, you can even start with the default styles and tweak them to suit your brand.

If you’ve got a bit of budget (even under €5K), consider working with a free- lance designer. There are plenty of talented ones who offer high-quality work at reasonable rates. A good freelancer can help you quickly pull together:

• A simple but professional logo

• A basic graphic system (font, color, layout rules)

• A lightweight style guide or Figma kit that makes it easy to apply consistently

Let’s walk through the essential visual brand touchpoints, the places where your design shows up.

Working with freelancers

If you manage to DIY everything, that’s amazing. But if you’re struggling and you have a small budget to spare, you should consider working with a freelancer or a small studio agency. Keep the large agencies for later on as you scale and need more strategic work.

Finding the right freelancers is about smart networking. Start by asking other founders, especially those with brands you admire, who they worked with. Most are happy to share their contacts, and this comes with the added benefit of a built-in reference.

Online platforms like Dribbble and Behance are gold mines for finding designers with specific aesthetics.

Once you’ve found someone promising, the success of the project hinges on your communication. Create a simple brief that clearly outlines what you need, when you need it, and why it matters to your company. Share the brand foundations and feeling you’ve already worked on - this strategic work will save them a lot of time, and you a lot of money.

Provide concrete reference points by sharing 3-5 examples of brands that capture elements of what you’re looking for, with specific callouts of what you like about each. “We love how this feels approachable but still technical” is much more helpful than “make it modern.”

Bringing your brand to life - Comms

Defining your core message

Earlier in the chapter, you defined your brand elements - your vision, mission, benefits, and what drives you. That work gives you internal clarity. But when it comes to communicating with the outside world, you need a core message.

It’s not a tagline or a sentence to slap on every slide or web page, but the thread that runs through all your comms. It’s the one idea you want people to walk away with, whether they’re a customer, candidate, investor, or friend of a friend.

It can show up in many different words, but it almost always shows up in spirit.

A strong core message communicates:

1. Who you are
2. Why you’re different
3. Your brand feeling

You’ll refine this message over time, and you might adapt it for different au- diences. But early on, one clear message is enough.

Examples:

Spendesk

“Where finance connects” – Reframes Spendesk as more than a tool, it’s the connective layer between finance teams, employees, and spending decisions. It speaks to trust and collaboration in one elegant line.

Swan

“Banking features that feel like product, not infrastructure” – Captures a strategic shift: finance is no longer just a backend layer, it’s something your users experience directly. Swan enables companies to embed branded fi- nancial features into their product quickly and seamlessly, making embedded finance part of your value proposition, not your tech stack.

Aircall

“The only phone system that feels like your other SaaS tools” – Aircall behaves the way modern teams expect software to behave: fast to set up, easy to use, and built to work seamlessly with the rest of your stack. This positions it as the phone system designed for the SaaS generation, not a relic from the telecom past.

Getting your message across

Once you’ve defined the message you want to get across, the next challenge is making sure they actually hear it and hear it in the right places.

You don’t need to be on every channel. In fact, spreading yourself too thin early on usually backfires. What matters most is being intentional: showing up where your audience already spends time, in formats they naturally engage with.

One common mistake people make is starting with the channel. You de- cide you want to launch a podcast, or start a blog, or write a newsletter then scramble to find something to say. That approach is backwards.

Instead, start by defining the message first, that way, you can make a more intentional decision about which channels are best suited to deliver it, whether that’s:

• Website

• Social media

• Blog

• Newsletter

• Podcasting

• Youtube / video

• PR

• Live demos

• Events

• ...

Not every message belongs on every channel. Some things are better suited to quick, skimmable formats like LinkedIn. Others need longer-form content like a blog post, podcast, or founder letter. The key is matching your message to the format that will carry it best.

And don’t overcomplicate it. Early on, you only need one or two channels. Pick the ones where your audience is most active and where you’ll actually follow through.

Channel spotlight: your website

Your website is one of the most powerful places to communicate your key messages. It’s often the first (and sometimes only) place where users, candidates, or investors will encounter your company, which makes it your single most important brand touchpoint.

Start with your tagline

Your tagline is usually the headline on your landing page, the first thing people see. It should clearly express the core benefit of your product (refer back to your ‘brand elements’ doc to find it) while reflecting your brand feeling. Here’s what makes a good early-stage tagline:

• Clear: no jargon or clever wordplay that hides the point

• Benefit-led: focus on what the user gets, not what you’ve built

• Emotional tone: calm, bold, empowering, urgent — whatever fits your brand feeling

• Short: aim for 5-8 words

Structure to follow

Here’s a suggested structure to follow that can be adapted to the specificities of your business.

• Hero: Your tagline, 1–2 supporting lines, and a clear call to action

• Social proof: Logos, testimonials, or early traction to build credibility

• Benefits: The key benefits of/use cases for using your product, e.g. “proto- type quickly”

• Key features: 3–4 features with light visual support (icons, screenshots) How it works:

• A simple 3-step overview of the user journey

• Call to action: Repeat your primary CTA

• FAQ: Pre-empt questions and remove friction to conversion

Tips to keep in mind

• Don’t try to say everything - focus on what converts best

• Lead with outcomes – your headline should tell people what they’ll gain

• Write for skimmers – use clear headers, bullets, and short paragraphs Prioritise benefits over features – talk about them, not you

• Stay aligned with your brand feeling – what emotion do you want to provoke in people?

• Keep it human – use simple, confident language

• Be ruthless – only include what helps someone say yes

Channel spotlight: founder branding

In the early days of your startup, your personal brand as a founder is often your most powerful marketing channel. While your company is still unknown, you are immediately visible and accessible.

Your personal presence is the approachable entry point that draws people in before they fully understand your product.

Align with your brand’s feeling, but stay authentic

Your personal brand should exist in the same universe as your company’s brand, but it doesn’t need to be identical. If your company’s feeling is about empowerment, you don’t need to become a caricature of empowerment in your personal presence.

The most effective approach is to find the natural overlap between your authentic self and your company’s values. Your personal content might be more casual, more opinionated, or more playful than your company’s voice - and that’s perfectly fine.

Tell your story

Your personal story (why you started the company and what problem you experienced) is one of your most valuable brand assets. It humanizes your company and creates connections in ways that talking about your product can’t.

To tap into your founder story in a way that aligns with your brand:

• Identify the emotional core: What feeling drove you to start this company? Frustration? Inspiration? Determination? This should connect to your desired brand feeling.

• Find the relatable moments: Which parts of your journey will resonate most with your audience’s own experiences?

• Connect past to present: Show how your personal history directly shaped your company’s approach or values.

• Keep it authentic: Share genuine challenges alongside victories, perfect stories are less believable and relatable.

Remember that you’re building two parallel but connected brands - your personal reputation and your company’s identity. When managed deliberately, they strengthen each other, creating more entry points for people to discover and connect with your vision.

Channel spotlight: PR

PR is obviously an important lever in any GTM strategy, particularly when it comes to driving inbound interest and creating early momentum around your product. In that sense, this section could easily sit within the GTM chapter. However, PR also plays a broader, longer-term role. It’s not just about short- term traction, but about shaping perception, building credibility, and increasing brand awareness over time. PR is especially valuable in the early stages of

a startup, when you’re still building trust. A well-placed article or feature in a reputable publication can act as powerful social proof. At a time when no one has heard of you yet, earned media can help reinforce what your brand stands for, and bring that story into circles you wouldn’t reach on your own.

Be genuinely helpful

Start by identifying journalists who regularly cover your industry. Read their articles, understand their angles, and get familiar with the topics they care about. If you notice gaps in their reporting or have insights that could deepen their coverage, don’t hesitate to reach out—share constructive feedback, offer data, or connect them with experts in your network.

The goal is to build a relationship before you need something. Position your- self as an expert voice in the space. When the time comes to pitch your own news, they’ll already know who you are and how you think about the industry. The most effective relationships with the press are built on mutual value, not transactional asks.

How to get them interested

When you’ve got something to announce and want press coverage, the first question to ask yourself is: is this truly newsworthy? Journalists aren’t going to write about your first sales hire. And unless you’re a tiny startup landing a major, high-profile customer, that alone likely won’t make the cut either. What might get coverage? Things like a significant product launch, a noteworthy fundraising round, or traction that signals a broader trend. But even then, it has to be positioned well and actually stand out.

Also, remember: journalists aren’t here to provide free advertising for your startup, they’re here to write stories that matter to their audience. That’s why it’s crucial to understand who that audience is and tailor your pitch accordingly.

Here are a few tips to help cut through the noise and connect with journalists who receive hundreds of pitches every day:

• Uniqueness - What makes your news genuinely new or different from anything else out there? Why are you uniquely positioned to solve this problem better than anyone else?

• Relevance - How does your story tie into a broader industry shift or cultural trend? Go beyond what your product does for your users, frame it in terms of a growing problem or unmet need that many people or companies are starting to face.

• Humility - Avoid buzzwords and vague claims. Especially in a climate saturated with AI jargon and overblown promises. Journalists can spot fluff from a mile away, make every word count.

• Expertise - Demonstrate a strong grasp of your industry. If you can position yourself as a credible voice (not just about your product, but about the space as a whole) journalists will start to see you as a go-to expert. That means they’re more likely to come back to you for quotes or background on future stories.

Keep in mind: a funding announcement alone might not be enough to spark interest. What’s unique about your investors, your timing, or your market? Are you solving a pressing, widely felt problem in a new way? Are you riding, or resisting, a major trend? The sharper your angle, the better your chances of getting picked up.

Example: Roundtable

Rather than simply announcing their €3M seed round, Roundtable crafted a compelling narrative: “100 business angels from across Europe back a new way to fund startups.” This angle aligned perfectly with their brand mission and tapped into ongoing discussions about early-stage investing. Their clear messaging secured substantial coverage across major tier-one publications.

Be clear on your angle

Imagine this: if you could choose the headline of the TechCrunch article that features your startup tomorrow, what would it be?

That’s the level of clarity you need. Your ideal headline should guide how you craft your press release and how you pitch your story. It should appear in your subject line, your opening paragraph, and be echoed throughout your messaging.

Journalists ultimately decide how they frame a story but if you’re unclear about your positioning, you leave that framing entirely in their hands. And that can lead to a message that misses the mark or fails to resonate.

The clearer and more compelling your angle, the more likely it is to be picked up and accurately represented.

Pitch yourself

Early on, you don’t need a PR agency. In fact, most agencies aren’t a good fit before Series A, they’re costly. At this stage, you don’t need polish, it’s best for you to start building relationships and pitch to journalists.

Start with:

1. A short, authentic pitch

Write a concise email:

• Clearly lead with your news and why it matters.

• Explain your product simply, in one or two sentences.

• Include a link to the press release in your email (more details below).

• Zoom out: Why is this relevant right now? What’s the bigger shift you’re part of?

2. A thoughtful target list

• Look at coverage your competitors received. Who covered them?

• Read recent pieces from those journalists. Understand what resonates.

• Prioritize media your ideal audience genuinely reads, including industry-specific outlets.

3. Personalize every email

• Mention something specific the journalist wrote.

• Keep it brief, conversational, and buzzword-free.

• Offer interviews, data, or insights that genuinely add value to their readers.

Sending 5–10 thoughtful emails is infinitely more effective than blasting out dozens of generic pitches.

Make journalists’ lives easier

Journalists are juggling tight deadlines and overflowing inboxes, so the less friction you create, the more likely you are to get coverage.

Here’s how:

• Prepare a press release (a Notion doc works fine) that’s easy to reference and written in the right tone:

  • Keep it neutral and factual — Press releases must be written in the third person, using clear, impartial language. Avoid opinions, feelings, or hype.

  • Use quotes to express opinions — If you want to include emotion or pers- pective, do it through quotes (typically from the founder, a customer, or an investor). Keep it to 2–4 concise, impactful quotes max.

  • Lead with the essentials — Start with the most important facts: who, what, when, where, and why. Keep the entire release to one page with a strong, clear headline.

• Include a media kit with all the assets a journalist might need: founder bios, product screenshots, company logo files, key facts and figures, customer testimonials, and anything else that makes their job easier.

• Be available and responsive if they follow up with questions or request an interview. Timeliness matters.

• Don’t micromanage their work once the story is published, only reach out to correct factual inaccuracies. Respect their editorial independence.

For high-priority announcements, consider offering an embargo - a way for journalists to prepare their piece ahead of time, with a clear publication date. It shows you understand their workflow and builds professional trust.

When to get help

For early-stage founders, hiring a full-scale PR agency is rarely worth the cost. But working with a freelance PR consultant can be a smart move, especially in situations like these:

• You have no idea where to start, you’re short on time, have no media contacts, and you need someone to maximize the impact of your first big announcement.

• You’re getting a sudden surge of journalist interest, with lots of inbound requests, and you’re unsure how to prioritize or respond strategically.

• You’ve landed opportunities for live interviews (TV, radio, panels) and need media training to present confidently and stay on message.

• You’re planning an international announcement and need help localizing your press release or connecting with media in other regions.

That said, always stay actively involved in your messaging and outreach. No one knows your story better than you do, and journalists often prefer hearing directly from founders. A PR freelancer can help you refine your story, manage logistics, and open doors—but don’t outsource the relationship-building. Taking the time to build direct connections with journalists is one of the most valuable long-term PR investments you can make.

The one question that matters

Whether you’re pitching a journalist, designing your landing page, or creating content, the success of your brand ultimately hinges on one question:

“What do I want people to feel when they encounter my company?”

This single question cuts through the complexity, turning abstract concepts into concrete choices. So as you build your brand in these early days, keep coming back to this question. Let it guide your choices, big and small. Your answer might evolve as you learn and grow, but the question itself will always be your most valuable brand asset.

To get the Hexa Roundup delivered every month to your inbox, don’t forget to subscribe below - or go to your account settings to turn of Hexa Roundup emails.

Subscribe now

🎙️ PMF Collector with Rodolphe Ardant (Spendesk)

Rodolphe didn’t start with a polished product. He started with a browser extension that didn’t work. Then he found a partner, issued virtual cards, and watched usage grow week over week at every client. In this episode of PMF Collector hosted by Mat Vaxelaire, the Spendesk cofounder covers why your first MVP failing can still point you toward the right signal, the 3 concrete moments that told him he had a real business, how to raise a round in 3 weeks after 3 months of VC rejection - and why ignoring your instincts in favour of outside advice is almost always a mistake.

🛡️ Europe’s defense tech moment - a collaboration with Mind the Gap

Hexa Sprint’s Emma Caillat and Agathe Thiery (founder, Mind the Gap) co-authored a piece on European technologies driving real geopolitical and societal impact. Hexa Sprint is currently supporting defense startups through its program - €500K investment and operational support. If you’re building in defense, applications are open.

⚡ Event: CTO Voices x OpenAI

OpenAI has shipped a lot lately: Super App, Codex, 111 new plugins, GPT-Rosalind, a new Agents SDK, ChatGPT Images 2.0, and $122B raised along the way. This Wednesday, Katia Gil Guzman (formerly CTO at Emera, now at OpenAI) joins Hexa CTO Tanguy Goretti for a conversation on how to get the most out of these tools without getting caught in the hype cycle. Drinks follow.

→ Register for the event

Portfolio news

🌱 Oligoplus (HX26) joins Hexa Sprint

Maxime Imhoff and the Oligoplus team join Hexa Sprint’s Climate cohort. Their bet: precision foliar nutrition by drone. Agronomists identify deficiencies, apply the right micronutrients at the right moment. Farmers pay for results — not inputs. That means +30 to +50% yield increases, less dependency on synthetic fertilizers, and healthier crops. €1.5M revenue in 2025, fully bootstrapped. Target: €5M in 2026.

→ Learn more

🏦 Marble (HX22) launches v1: autonomous AML & fraud monitoring

Marble v1 is live. The platform now covers the full compliance journey (from screening and transaction monitoring to automated detection and AI-assisted investigations) with no gaps between modules. A swarm of AI agents handles what the industry has been billing for separately: tuning detection rules, calibrating scoring models, pre-qualifying alerts. All under the direct supervision of the compliance officer, not a vendor. 100+ financial institutions from the US to Singapore are already on it. Open-source core. On-prem or SaaS.

→ See what’s new

🥞 Pancake (ex-Basalt - HX24) — new name, beta next week

Basalt is now Pancake. The team has been heads-down on something big: new name, logo, and identity: done. 50 alpha users already testing: done. Beta launch: next week. More news coming very soon.

→ Stay tuned

Hexa news

🚀 Six new ideas live on hexa.com/start

Hexa Start Partners Florent Quinti, Mat Vaxelaire, and Raphaël Vullierme each bring their own ideas to the table and are looking for the right founders to build them together. Six highly ambitious bets: QAx, CompanionX, LLMx, PhysicalAIx, CustomsX, PayX. Twelve months of hands-on support, a team to help you build and close your first customers, and three months fully covered in San Francisco.

→ Explore the ideas at hexa.com/start

🤝 Hexa partners with OpenAI

Every startup in the Hexa portfolio now gets direct access to OpenAI from day one: credits, model access, shared engineering support, and co-hosted events. At Hexa, AI is the core architecture of every company we build, not a feature. From Parallel deploying agents inside hospital workflows to Waniwani helping companies sell inside AI assistants, our founders are reimagining industries from the ground up. This partnership means they’ll have the best tools to do it.

→ Learn more

🏛️ Inside Hexa: investor events in Brussels & Paris

We opened our doors to LPs and investors to share what’s changed. The conversations covered where value is being created, which business models will be reshaped, and where capital should flow next. Hexa Start, Hexa Sprint, and Hexa Scale each presented, with founders from Panora, Forest Admin, Spendesk, Enobase, and Yousign joining to give their perspective.

Thanks for reading! Follow us on Linkedin and Instagram to stay up to date between issues.

Cheers,

The Hexa Team

This article is written by (founder at Mind the Gap) & (Defense Tech & European Sovereignty - VC @ Hexa Sprint). It is a collaboration between Mind the gap and Hexa Sprint. Mind The Gap investigates European technologies and innovations driving real geopolitical and societal impact.

Hexa recently launched a new accelerator, Hexa Sprint, which is currently supporting defense startups through its program (500k investment and operational support)

If you’re building in defense and looking to accelerate your growth, you can apply here.

When we first met at Hexa on a Sunday afternoon a few months ago, we had no idea we would end up here. Friday evening. 6:30 PM. Martin de Gourcuff welcomes us warmly to his bunker-like offices and shows us around the former production lines of the startup turned unicorn.

The next day, the news is brutal: the war in Iran has just begun. Like the Russian-Ukrainian war before it, this new conflict reveals a profound paradigm shift, a new era of the battlefield. While the collective imagination still pictures the so-called “Primes”, like Safran, Thales, or Airbus, as the drivers of innovation, it is now startups that are developing cutting-edge technological products at record speed, addressing the challenges of today’s conflicts directly on the ground. These startups support the military directly on the front lines, delivering the best solutions in a world where everything moves fast. A world where mechanization gives way to dronization, where automation and artificial intelligence play an ever-growing role, and where electronic warfare has become a major stake in modern conflicts.

Beyond software, it is also production capacity, factories, and supply chains that are changing the game, at a time when the multiplication of regional conflicts makes a large-scale war increasingly plausible. At the same time, an existential question arises: is Europe capable of producing its own weapons, on its own soil, with full control over its components?

It is in this context that we had the chance to meet Martin de Gourcuff, co-founder and CTO of Harmattan AI, but also Hadrien Canter, CEO of Alta Ares, and Arnaud Valli, Head of Public Affairs at Comand AI. Come on board with us to the heart of the “New Defense Edge” to discover these French startups at the forefront of modern conflicts.

Part 1: New rules, New players, New game

The peace dividend era is over. As global conflict intensifies and the shadow of a large-scale war has never loomed closer, Europe no longer has a choice: the time has come to rearm.

Asymmetric wars have multiplied and complicated threats (electronic warfare, drone swarms, cyber attacks), creating new challenges that traditional defense strategies struggle to address. The Russo-Ukrainian conflict has provided a brutal demonstration: the most sophisticated weapons no longer guarantee victory. Drones costing a few thousand euros are destroying armored vehicles worth several millions. To quote Andrius Kubilius, European Commissioner for Defense and Space, we are now witnessing “€1M missiles facing €10,000 drones.” The conclusion is clear: the rules of the game have changed.

European defense is going through a pivotal moment, shaped by both geopolitical tensions and economic transformations. Under the constant pressure of an open war at its doorstep, and caught in the grip of Sino-American rivalry, Europe must respond. In this new paradigm, traditional defense players are less equipped to operate at the required pace, leaving considerable room for startups capable of iterating quickly and adapting to constantly evolving threats.

Harmattan AI, Alta Ares, Comand AI: the new French startups redefining the European defense landscape.

Born from this urgency to rearm, these three companies were built around the same conviction: today’s threats call for responses that traditional players cannot provide.

Cheap to produce, practically invisible to traditional radars, and capable of being deployed in massive swarms, drones create a radical asymmetry against classic interception systems, which they overwhelm as easily as they bypass. This is the problem Harmattan AI, founded in April 2024, is tackling. Initially focused on detecting anti-personnel mines (an activity it continues to develop in parallel), the startup gradually focused on drone interception, offering verticalized solutions from software to hardware. A trajectory that convinces investors: in January 2026, Harmattan raised $200 million, bringing its valuation to $1.4 billion.

Alta Ares takes a similar path, but its origin is radically different. Founded two years ago directly in Ukraine by Hadrien Canter, the startup was literally born on the front line. Its first solution, Gamma ISR, helps relieve the cognitive load of operators forced to scan video feeds for hours. The second, Pixel Lock C-UAS, powered by artificial intelligence, allows for the automatic interception of loitering munitions like Shahed drones. Validated on the Ukrainian battlefield, the technology has since been tested under extreme conditions in Estonia and validated by NATO.

Beyond counter-drone warfare, modern conflict poses another equally critical challenge: command. In environments saturated with sensors, drones, cyber signals, and OSINT, decision-makers face unprecedented volumes of data, well beyond what a human alone can process in real time. This is precisely the challenge Comand AI wanted to address. Founded by Loïc Mougeolle, former Director of Strategic Innovation at Naval Group, the startup develops military command software enabling faster decision-making by officers, along with automated after-action reviews to extract real-time lessons from operations.

Three companies, three approaches, but a shared genesis. That of a generation that chose to put its skills in service of European defense, not out of opportunism, but out of conviction. Arnaud Valli, former political advisor to NATO and now Head of Public Affairs at Comand AI, summarizes this shift: “We are right in the middle of an AI arms race, and if we are going to have effective systems, it would be better if liberal democracies had them. And I agreed with the secondary option: it would be good if Europe and France had them too.”

A sentiment shared by Martin de Gourcuff, who admits he never imagined working on defense systems: “Today, I have a much clearer sense of the purpose it serves.” At Alta Ares, this utility takes an even more concrete form. “Alta Ares is a mission-driven company,” asserts Hadrien Canter. “There are few companies today in Europe that code in the morning, and by the evening, what they coded is running on interceptor drones shooting down Shaheds. And that is a point of pride.”

Software as a lever for operational superiority

War has always been a technological accelerator. From gunpowder to the atomic bomb, from World War II radars to the drones of Mosul, every major conflict has reconfigured the relationship between innovation and military power. But something has fundamentally changed this time: the decisive advantage is no longer about tonnage or caliber. It comes down to the software layer, the one that turns an ordinary drone into a precision system, or a thirty-year-old platform into an augmented combat tool.

Beyond the robotization of the battlefield, we are now witnessing the autonomization of the entire chain, from decision-making to combat action. As Martin de Gourcuff sums it up: “We can add far more features to the defense systems we develop thanks to software and algorithmic advances, especially in machine learning. And that, in particular, makes it possible to render the systems much more autonomous than they were before.”

In a context where defense budgets do not allow everything to be replaced, and where acquisition cycles remain long, software becomes the fastest and most cost-effective lever to increase the power of existing arsenals. Software thus improves hardware, as Arnaud Valli points out: “We have systems like the T55, a tank dating back to the 50s, which is still used but is improved thanks to software. Software is not just an application overlay: it is a way to improve the hardware, to make it more competent and more effective [...]. Software is still quite decisive in today’s digitized armies, which are increasingly robotic.”

A paradigm shift in the defense industry

In a war of position like in Ukraine, this autonomization translates into a concrete operational advantage: mobilizing fewer soldiers, in a context where armies are struggling to recruit, with better-performing equipment. Above all, it provides an agility that large industrial groups are structurally incapable of offering. In Ukraine, the rules of the game evolve in a matter of weeks: jamming, fiber-optic drones, electronic countermeasures, new flight profiles. The Primes, excellent at building sophisticated systems over long cycles, find themselves outpaced by the rhythm of a war of attrition where what matters most is resilience against enemy innovation and the ability to mass-produce.

Arnaud Valli, who observed this system from the inside during his years at NATO, points out its limitations: “In France, the armed forces make a request, the DGA translates this operational need into an industrial need and asks the manufacturers to produce it. And the user hardly ever spoke to the manufacturer, except upon delivery. Which, unfortunately, sometimes resulted in absurdities between the demands of the operators and the final product.” Startups, on the other hand, operate with a different mode of functioning, often referred to as DefTech. “What does that imply? It is a much more iterative way of working, directly with the users,” specifies Arnaud Valli.

Major industrial players have clearly taken the measure of these new actors: Dassault Aviation took a stake in Harmattan AI, while Safran acquired Preligens, a French startup using AI for satellite image analysis. These moves go beyond mere financial logic: they reflect an assumed complementarity. Hadrien Canter places this dynamic in its broader context: “I think we are right in this pre-war phase, and that we have an enormous responsibility, which obliges us to have the greatest rigor in what we do. France has an important role to play in building the European defense of tomorrow. And this role can become a reality with the help of startups, but also with the help of large corporate groups.”

The awakening of institutions and private players

This realization is not limited to large manufacturers; institutions are also beginning their shift. Certain initiatives have emerged, such as the creation of the Future Combat Command (CCF) or the Defense Innovation Agency (AID), aimed at modernizing the French army and integrating startups into the Defense Technological and Industrial Base (BITD). Hadrien Canter measures its daily impact: “There is a real awareness compared to when we started two years ago. This is largely due to an organization called the Drone Pact. It is a real success; it helps us enormously to identify the right points of contact and to bridge the gap between the final need and the manufacturer. It is a link that was stretched too thin for too long, and we take great pleasure in shortening it.”

This openness also translates into concrete contracts: Harmattan AI has secured commitments with the French and British armies, a strong signal of institutional credibility acquired in barely two years of existence.

Private capital is following suit. Funds like Sisyphus Ventures or Expansion, and dedicated acceleration programs like Hexa Sprint Defense, have turned DefTech into a full-fledged asset class, where it was still taboo a few years ago. Martin de Gourcuff observes this in his own fundraising: “There has been a huge shift. Investors are much more open now to the defense industry. Many of them told us that we were the first defense company they put money into.”

The BITD is transforming. It attracts capital, talent, and a new generation of founders who would not have imagined building weapons systems five years ago. The cards are being reshuffled, and for Europe, the momentum is here. The challenge now is to transform it into the foundations of lasting sovereignty.

Partie 2 : Produce or Perish

Sovereignty is no longer an abstract concept. In a world where armed conflict is once again the norm, it refers to a concrete and brutal reality: the ability of a State to decide its own fate on its own territory, without depending on the goodwill of an ally or the supply chain of an adversary. This is exactly what Ukraine has been fighting for since February 2022, and it is precisely where the three startups we spoke with have chosen to commit themselves.

Because being “on the edge” is not a metaphor. The solutions from Alta Ares, Harmattan AI, and Comand AI are not tested in Parisian labs or German simulation centers; they are running on the front line. Alta Ares produces its interceptor drones in Ukraine, as close as possible to the units that use them. Harmattan AI and Comand AI are also deployed there. These startups do not use war as an experimental playground: they respond to real, immediate, and vital needs. Every successful interception is an impact avoided. Hadrien Canter testifies to this: “Today the results are quite good, we have already been able to shoot down more than a dozen Shaheds. And we keep counting. Every night. And we tell ourselves that every Shahed shot down is potentially a piece of critical infrastructure that is protected.”

Beyond their agility as startups and their software-oriented product design strategy, the context of high-intensity war on Europe’s doorstep has allowed these companies to test their solutions in real conditions and produce for immediate needs. “One week of development for us in Ukraine is the equivalent of two months in Germany or France,” Arnaud Valli emphasizes. This access to a massive market hungry for innovative options and products has allowed them to develop solutions that meet the needs of the Ukrainian armies for their fight, for their freedom. And more broadly, these solutions also allow Europe and European armies to prepare for future battles.

The situation in the Middle East has reinforced, with added brutality, what Ukraine had already demonstrated. Iran was one of the first States to systematize the large-scale use of low-cost drones, the Shaheds, produced for a few thousand euros, capable of inflicting considerable damage, and which cost several millions to intercept. A radical economic asymmetry, revealing a fundamental rule ignored for too long: in a war of attrition, mastering production capabilities is just as decisive as the sophistication of the systems.

The rapid depletion of missile stockpiles in the region is the starkest illustration. And while some European leaders have become aware of this reality, translating it into concrete industrial capabilities remains painfully slow. As Hadrien Canter observes: “Our adversaries understood long ago that quantity has become a quality. And unfortunately, some of our leaders have understood it, but it takes time to implement.”

Designing high-performance and autonomous systems is one thing. Producing them on a large scale, at a low cost, and without depending on foreign powers is quite another, and this is often where everything is decided. In a war of attrition, the supply chain is not a logistical detail: it is a full-fledged strategic issue. The factory has become a weapon once again. And mastering it is a condition for sovereignty.

Whoever controls their supply chain controls their destiny

This change of scale shatters Western military orthodoxy. Until now, calibrated for one-off expeditionary operations, our armies favored ultra-sophisticated systems produced in very small numbers. But once a conflict turns into a war of position, the real moat inevitably moves to the assembly line.

Survival no longer depends solely on R&D, but on the absolute primacy of physical production capacity. A brutal reality summed up by Martin de Gourcuff (Harmattan):

“In Ukraine, what makes the difference is being able to produce hundreds of thousands of units. It is not so much the intrinsic performance of the system, although obviously that helps. But performance and cost are generally highly correlated; so we end up with a fundamental trade-off to make between performance and mass.”

In a world where high intensity is becoming the norm again, mass production implies total control of the supply chain. For our Defense Tech startups, the ambition is clear, albeit complex: breaking free from the double dependency on China and the United States.

The case of the United States is particularly tricky. While they remain historical allies, protectionist tensions, exacerbated by Donald Trump’s return to the White House, are rewriting the rules. The main threat to European companies lies in the ITAR (International Traffic in Arms Regulations) compliance. The moment a single American component subject to ITAR is integrated into a system, no matter how small, the entire final product is “infected” by the standard, forcing the European company to request an export license from the US government.

On the software side, the market remains largely dominated by American tech giants. To train its models, Comand AI must source GPUs outside of China, turning to Taiwan, South Korea, or Vietnam. At Harmattan AI, which uses powerful hardware accelerators, there is currently no viable alternative to the American giant Nvidia.

At Alta Ares, strategic partnerships in Asia, particularly with Taiwan, are a deliberate choice, as Hadrien Canter explains:

“We work with Taiwan for production capacities and niche expertise, but also because they are an ally. Taiwan faces the same threat as Ukraine today, and buys French equipment. We created this company out of conviction. When we know that China supplies 75% of the electronic components for Iranian Shahed drones, why shouldn’t we provide our assistance to Taiwan?”

Wanting to relocate the entire value chain back to Europe, however, runs up against the wall of economic realities. The production of cutting-edge semiconductors requires massive investments and relies on indispensable civilian players. Arnaud Valli (Comand AI) provides a lucid assessment:

“We are dependent on civilian market structures to get our GPUs. As long as Nvidia manufactures its cards in Asia, we will be dependent. The smallest TSMC factory in the United States requires $12 billion in investment. Relocating therefore requires a strategic investment vision from the State, from data centers to production, and above all accepting, at least for a time, to produce at a loss compared to components made in Vietnam that will always be cheaper. It is a real political and industrial choice.”

Furthermore, the chain is globalized and hyper-specialized: TSMC cannot manufacture its chips without extreme ultraviolet (EUV) lithography machines from the Dutch company ASML, which holds a near-global monopoly on this technology. Europe cannot do without Taiwan, which cannot do without Europe.

While there are vast European programs to relocate the manufacturing of micro-components or batteries, the road is long. In the meantime, the hardware remains dependent on Asian or American civilian markets for cost and labor reasons. Faced with China’s hegemony over rare earths, which are essential for motor magnets, batteries, or radiofrequency systems like radars, agility is essential. French startups are expanding internationally to secure their supplies and markets: Alta Ares is forging ties in the Middle East for radar and electro-optical expertise, while Harmattan AI is expanding into Switzerland, the United States, and the United Arab Emirates.

European Defense: An unfinished puzzle

Ultimately, building a sovereign supply chain raises a much broader question: the harmonization of European defense systems. And the task is colossal.

Despite their geographical proximity, European countries think about war through the prism of their own history. A former colonial power does not project its forces in the same way as Germany, which was militarily constrained for a long time, or Poland, which lives with the Russian threat at its borders. This divergence of objectives creates an industrial fragmentation bordering on the absurd: “In Europe, there are 15 different types of tanks. In the United States, there is only one,” points out Martin de Gourcuff.

The relationship with the American nuclear umbrella is just as divisive. Driven by de Gaulle’s vision, France built a strategic autonomy unique in Europe, distinguishing itself by its ability to design its own aircraft carriers, nuclear submarines, and fighter jets. Constrained by a small domestic market, this model contrasts with the reality of many European neighbors, strategically and technologically vassalized by Washington.

One might have thought that recent geopolitical upheavals, up to Donald Trump’s most surreal proposals, like his idea of taking Greenland back from Denmark (by buying it or, if necessary, by force),would be a definitive wake-up call. But the inertia remains, and many Europeans still refuse to consider that their historical ally might prioritize its own interests to the detriment of those of the Old Continent. Faced with these industrial and political realities, one question remains unanswered: is European Defense condemned to be nothing more than a chimera?

Part 3: Hardcoding the Red Line

The sovereignty imperative goes far beyond simply mastering hardware components. Today, the true frontline lies in the adoption of artificial intelligence within our defense systems. While the technical and strategic challenges are vast, the most pressing questions remain ethical in nature and fundamentally entangled with the growing autonomy of our arsenals.

Many whistleblowers are raising concerns about the rapid development of AI. Could an artificial intelligence with amplified capabilities become a systemic threat to humanity? While these narratives echo the most famous works of science fiction, the concerns are real and voiced by leading figures in the ecosystem itself. In a recent essay, Dario Amodei, CEO of Anthropic, clearly outlines the threats posed by AI if it is militarized without safeguards: the risk of a hostile nation achieving military dominance or launching large-scale cyberattacks, the appropriation of these tools by terrorist groups to multiply their destructive capacity, or the concentration of absolute power in the hands of dictators.

Described as “patriotic” by Arnaud Valli, this essay is a strike warning against the potential excesses of defense AI, but also raises deep philosophical questions. Martin de Gourcuff shares that wariness, and confirms that there is no such thing as zero-risk when dealing with fully autonomous systems:

“LLMs provide access to a vast body of knowledge and open up a wide range of possibilities. Unfortunately, it is a technology that can be used for harmful purposes. How worried should we be? I’m not entirely sure. But we must not forget that this tool is not fundamentally constrained to do good.”

In this context, a key question emerges: does the growing implementation of AI in Defense inherently represent a risk?

The startup paradox: more AI for more ethics?

When it comes to weaponry, AI primarily translates into autonomy. While automated systems (drones, missiles) have existed for decades, the current robotization of the battlefield marks a new step forward. Yet for our startups, the conclusion is unanimous and almost counterintuitive: controlled automation enables a level of efficiency and performance that strengthens the ethics of strikes.

In practice, this plays out in one concrete advantage: video feedback from autonomous drones makes strikes reversible. Where launching a traditional missile is irreversible, an interceptor drone strike can be canceled up to the final five seconds. Hadrien Canter recalls a tragedy that could have been avoided:

“When pro-Russian separatists and Russian armed forces shot down civilian flight MH17 over Ukraine in 2014, they believed they were targeting a Ukrainian military aircraft. [...] Today, when a drone is launched to intercept a target, if we realize that the radar lock was actually on a friendly or civilian drone, video feedback allows us to cancel the strike at the last second.”

Harmattan AI takes the argument one step further: specialization itself is the safeguard. “We are developing Gobi, a drone interceptor. Unless we modify its code, it can never be used to target civilians. From that perspective, it is almost more secure than a traditional anti-aircraft gun, which can easily be turned against anyone,” explains Martin de Gourcuff.

Automation is not limited to deploying swarms of drones; it relies on the resilience of communication networks to operate a global command and control (C2) system, similar to the American Combined JADC2 program. However, Arnaud Valli points out the fragility of this hyper-connectivity: “This connectivity can always be compromised, even if Starlink or other constellations make 5G available everywhere.” To what extent can armed forces rely on cloud availability in the middle of combat?

What makes this automation inevitable is a single imperative: accelerating the Kill Chain, the sequence of processing a target. The goal is to drastically reduce the time between detecting the enemy and neutralizing it.

“The Russians have improved their detection-recognition-strike system in a very impressive way: previously, it took them five minutes to identify and destroy a target, now it’s under 30 seconds,” notes Arnaud Valli.

AI is revolutionizing targeting, even reshaping the economics of war, as observed through the Iranian conflict. “During the first strikes [recent ones in the Middle East], there may have been 1,000 strikes managed by only around a hundred analysts… That is 100 times fewer personnel than during the Gulf War, for four times as many strikes. When applied to data aggregation, this multiplier effect is terrifying. But ultimately, it does not win a war on its own,” tempers Arnaud Valli.

The “Human-in-the-loop” and the wall of operational reality

Despite these advances, operational safety is only guaranteed if humans remain in the decision-making loop. AI helps reduce the operator’s cognitive load by filtering information. “AI provides suggestions, but humans remain at the core of the decision-making process,” insists Hadrien Canter.

Technically, autonomous defense systems face the same limitations as autonomous vehicles: handling rare events. Martin de Gourcuff illustrates this challenge:

“If a monkey crosses the road in front of our car, we understand the situation. The algorithm, however, does not: it either misidentifies the obstacle or makes a poor decision because its training data contains very few instances of monkeys crossing roads.”

These perceptual limitations, combined with sensor constraints (sunlight, hallucinations…), slow down the development of fully autonomous systems, even though technologies such as World Models aim to address them. Hadrien Canter sets the record straight regarding tech-driven fantasies: “While some competing companies oversell a futuristic vision where everything works at the push of a button… operational reality is very different. There is noise in the data, sensors get jammed. Human lives are at stake, and that must not be forgotten.”

Beyond the technical dimension, automation raises deeply political questions. Arnaud Valli highlights the risk of desensitization to war:

“It is clear that systems are becoming increasingly autonomous, with humans simply ‘turning the key’ because they can no longer compute everything. But if China were to deploy a fleet of robots in the Pacific against a US fleet, would there not be far less political and social resistance to initiating a conflict, knowing that there are no direct human casualties?”

The uncomfortable truth is that hostile powers aren’t waiting for a consensus on the ethics of autonomous warfare, they’re already fielding it, pushing Western armed forces toward an essential doctrinal renewal. Who, human or machine, is legally responsible for an autonomous strike? What delegated authorities to open fire are permitted? It is along this narrow tactical and moral backbone that Defense Tech must evolve. “Our companies must work with the military, but also with theorists, to build a clear, defined framework capable of evolving,” concludes Arnaud Valli.

Part 4: The war of tomorrow is being shaped today

As institutional trust strengthens and private capital finally opens up to Defense Tech, the time for mere proof of concept is over. For these startups, the ambition goes beyond providing isolated technological building blocks: the goal is to establish themselves as the backbone of tomorrow’s armed forces.

The ultimate objective? To become the central infrastructure, the global operating system of European defense. A direct response to American technological ambitions, such as the highly interconnected Combined JADC2 military program. Martin de Gourcuff confirms this desire to be at the core of the system: “There is real value in being at the heart of the data system. It is fully aligned with our ambition to build the operating system for command centers, the data infrastructure, and the communication systems of armed forces.”

To summarize this pursuit of interoperability and software dominance, Arnaud Valli offers a striking metaphor:

“Why do you think American Defense Tech giants use names drawn from Tolkien’s universe, like Palantir or Anduril? It’s one to rule them all. [...] In an ideal world, we would want a piece of software connected to everyone. The goal, within a European ecosystem, is to reach several billion in revenue. If we manage to shake up how procurement agencies operate in favor of end users, we will have achieved a great deal.”

This software sovereignty must also enable Europe to compensate for the shortcomings of its historical allies. Hadrien Canter sees this as a strong strategic signal: when the United States runs out of stock and can no longer supply Patriot missiles, the emergence of France and its ecosystem with agile and effective interception solutions shifts the balance on the global geopolitical chessboard.

But to turn this vision into reality and establish the continent’s security architecture, money and political will alone will not suffice. The true decisive factor of this decade will be human. “Join us to build concrete solutions and have an impact. There is work to be done, we are looking for motivated people committed to this mission-driven mindset,” says Hadrien Canter as a general call to action. The message is clear: tomorrow’s European sovereignty urgently needs the best talent of today.

Closing words

Writing this article opened our eyes to a world most of us instinctively turn away from: Defense, its mechanisms, its realities. A world opaque, sometimes uncomfortable but more urgent to understand than ever.

Because tomorrow’s threat may not yet have a name, but already exists. And while we look away, entire teams are coding, testing, iterating, sometimes just a few kilometers from the front line, so that Europe can defend itself in a world that did not wait for us to become hostile.

We thank Martin, Hadrien, and Arnaud for opening the doors of their offices to us. And we thank them even more now that we fully grasp the fight they carry every day, to protect the sovereignty of peoples. A fight that, ultimately, concerns us all.

If you enjoyed this edition, please consider sharing it with a colleague or friend who wants to understand how geopolitics and tech intertwine.

Best,

Agathe & Emma

Read up on the 10 takeaways from the episode:
1) The real problem wasn’t spending - it was how decisions were made

Before Spendesk existed, Rodolphe had already lived through the problem inside fast-growing companies. Teams were encouraged to move quickly and take ownership. That autonomy was essential to growth. But the moment money entered the equation, everything slowed down.

Finance, understandably, needed control. Visibility over cash, approval processes, compliance. The result was a system where every expense had to be validated and justified - often manually. And as the company scaled, that system became heavier and heavier.

At some point, the contradiction became obvious.

“I was rebuilding the exact bureaucracy I had hated before, just to regain control.”

That’s where the real insight came from. The issue wasn’t that employees were spending too much, or that finance lacked tools. It was that companies were trying to operate in a modern, decentralized way with systems designed for centralized control.

Spendesk was conceived as a way to resolve that structural mismatch: to give teams the ability to act, while allowing finance to keep a clear, real-time view of what was happening.

2) The first version failed and clarified everything

Like most early-stage products, the first version of Spendesk was far from what it would eventually become. The initial idea was to solve a very specific pain point: the sharing of company cards.

The solution they built was simple. A browser extension that allowed employees to access a shared company card, while logging who used it and when. It worked, technically speaking. But it didn’t solve the core issue.

Finance teams didn’t trust it. Security concerns remained. And most importantly, the product sat on the surface of the problem rather than addressing its root.

That failure turned out to be decisive. It forced a realization that would shape everything that followed: you can’t build a meaningful finance product if you don’t control the transaction itself.

In other words, visibility without control isn’t enough.

3) Building under constraints and learning faster because of it

At that stage, nothing was in place to make things easier. There were no obvious payment partners in Europe, no infrastructure readily available to issue cards, and no fully formed technical team to build the product quickly. Rodolphe himself went back to coding to push things forward, even though he hadn’t done it in years.

“I would spend four days building what a good engineer would do in one hour.”

From the outside, this looks inefficient. But in practice, it created something far more valuable than speed: proximity to the problem.

Eventually, they found a partner that allowed them to issue virtual cards. It wasn’t perfect, and the setup introduced friction that would normally kill adoption. But it was enough to test the next hypothesis.

4) The first real signal: customers taking irrational risks

The early product allowed companies to generate virtual cards for online purchases. To use it, however, customers had to pre-fund accounts - through a structure that was far from reassuring.

They had to wire money to Gibraltar.

And yet, customers did it.

That moment revealed something fundamental: the pain was strong enough that companies were willing to accept friction, uncertainty, and risk to solve it.

When users behave in ways that seem irrational on paper, it’s often because the problem you’re solving is much bigger than you thought.

5) The second signal: usage that grows instead of fading

If the first signal proved that the problem existed, the second showed that the solution was working.

Typically, early SaaS products see a burst of activity followed by decline. Users test, explore, and then slowly disengage when the product doesn’t integrate deeply enough into their workflows.

That didn’t happen here.

Week after week, usage increased. More transactions, more teams involved, more use cases emerging organically inside companies. What started as a tool for marketing teams quickly expanded to other departments, without any push from the product itself.

“Normally usage declines. For us, it kept increasing.”

6) The final signal: willingness to pay - without friction

Up to that point, everything had been free. Which meant one critical question remained unanswered: would companies actually pay for it?

When they introduced pricing, the expectation was cautious. Losing a significant portion of users would have been normal, even healthy. Instead, almost all customers converted.

Out of around a hundred companies, only a handful chose not to continue.

“We thought we would lose half. Almost everyone stayed.”

7) When investors don’t believe what customers already know

With these signals in hand, fundraising should have been straightforward. It wasn’t.

For several months, the response from investors was largely negative. The core idea — giving employees direct access to company spending — went against deeply ingrained assumptions about financial control.

“Companies will never give cards to employees.”

The breakthrough came when they changed their approach. Instead of trying to convince traditional venture capital, they turned to people who understood the problem directly: founders, operators, and early customers.

Within three weeks, the round was closed.

8) Building in Europe means building complexity from day one

Spendesk’s early development was rooted in Europe, largely because that’s where the founders had access to their first network and customers. But this choice came with a structural challenge that shaped the company’s trajectory.

Unlike the US, Europe is not a single market. Expanding means navigating multiple countries, each with its own regulations, behaviors, and operational constraints. What works in one market doesn’t automatically translate to another.

“You end up duplicating your effort in each country.”

This makes growth slower and more complex, but it also creates a different kind of strength. Companies that succeed in this environment build resilience early, and develop systems that can handle fragmentation.

9) The role of intuition when everything becomes less clear

Looking back, one of the most striking lessons doesn’t come from the early product decisions, but from how decisions evolve as the company grows.

In the beginning, founders rely heavily on instinct, simply because there are no established rules to follow. As the company scales, that instinct is often diluted by external input — advisors, investors, benchmarks, best practices.

That shift can be dangerous.

“Every time I ignored my intuition, it was a mistake.”

The challenge isn’t to reject advice, but to avoid overfitting to it. The ability to filter signal from noise becomes a critical skill, especially when navigating uncharted territory.

The takeaway

What makes Spendesk’s early story interesting isn’t just the outcome, but the sequence of signals that shaped it.

A problem that felt operational turned out to be structural. A failed MVP clarified what truly mattered. Customers validated the product not through feedback, but through behavior — by taking risks, increasing usage, and ultimately choosing to pay.

Behind it all, there’s a consistent pattern: progress didn’t come from perfect planning, but from continuously testing assumptions against reality, and adjusting quickly when they didn’t hold.

In the end, building a category isn’t about having the right idea from the start.

It’s about recognizing, early enough, when you’re onto something real and having the discipline to follow it all the way through.

In the early days of building a company, go-to-market is about experimentation. Your job isn’t to find the perfect channel or build the definitive playbook. It’s to test everything. Try ideas, kill what doesn’t work, double down on what does.

This is the part of the journey where nothing is obvious. You don’t yet know which channel will work best, what message will resonate, or which customer segment will bite first. And that’s normal.

Don’t fall into the trap of over-engineering your growth too early. It’s not the time to hire a performance marketing team or automate a funnel you haven’t validated. It’s a time for hands-on, founder-led discovery: testing outbound, inbound, partnerships, communities, content, even just good old conversations.

Think of this as your testing lab. The goal is to get from zero to something and to learn as much as you can in the process.

This chapter is your guide for that phase: how to test, what to test, and how to turn early signals into momentum that carries you toward your first 100 customers.

The founder’s role in go-to-market

In the early days, there’s no dedicated sales team. No growth hire. No account manager. There’s just you: the founder.

Before you delegate go-to-market (GTM) to experts, you need to understand the mechanics yourself. GTM at the start isn’t just about execution, it’s about experimentation and learning.

At this stage, a founder is the GTM team. That means wearing several hats:

• Sales: You need to know how to pitch, handle objections, close deals and understand why people say “yes” or “no.” If you want to attract great sales talent later, you need to prove you can sell first.

• Lead generation: You need to have an idea on how to find your early users, whether it’s cold outreach, warm networks, events, or content. One tip: be creative.

• Growth marketing: You need to test and learn quickly, from landing pages to email sequences to acquisition channels. What messaging gets attention? What drives conversion?

• Account management: You need to own the relationship after the sale. The best way to reduce churn is to stay close to your first users, solve their problems, and learn what “value” really means in their context.

But you won’t be good at everything. That’s normal. What matters is knowing where you’re strong and where you’re not.

Tomorrow, you’ll hire specialists to take over these roles. But today, your job is to master the basics and know where your strengths and weaknesses lie so you know exactly what kind of person to bring in next.

“In the beginning, like many entrepreneurs, we focused almost entirely on the product... But the problem is, even if your product is great, if nobody uses it, it’s worthless... From day one, think hard about how you’re going to distribute it. Search for your Product-Market Fit and your Go-To-Market Fit at the same time.”

– Alex Louisy, CEO and Cofounder of Upflow

“The goal of the early-stage GTM phase is to establish direct contact with the market and define your ICP. At this stage, you should avoid over-optimizing or automating too early. It’s important to not automate processes that you haven’t first validated manually. Especially now, with the explosion of AI tools and sales engagement platforms that make automation easier than ever.”

– Vasco Alexandre, CEO and Cofounder of Dotfile

Defining and refining your ICP

You won’t nail your ICP on day one - and that’s ok

A great ICP (Ideal Customer Profile) is narrow and precisely scoped. But getting there starts broad. Early on, you need to be open to different ICP possibilities and then test those hypotheses.

Begin by listing all the attributes that define your potential ICP(s). Your MVP and early discovery interviews will give you strong signals about who stands to benefit most. Use the list below as inspiration, not as a checklist to complete for the sake of “best practices.” Only include attributes that clearly align with the problem your product solves and help you concentrate your go-to-market efforts.

It’s normal at this stage to define not one, but two or three possible ICPs. That’s a good thing. Your job now is to describe them clearly so you can validate them later. The goal isn’t to land on the one right away, it’s to generate strong candidates to test.

Eventually, a solid ICP boils down to just a few key attributes. But early on, being too narrow too soon can be limiting - exploration comes first.

Here are some attributes to consider when defining your early ICP(s):

• Company size: instead of saying “SMB” or “mid-market,” get specific. If your product works best in companies with 100–300 employees because that’s when certain workflows start breaking down or when an IT director is in place, use that specific range.

• Sector: what verticals get the most value from your product? Healthcare? Fintech? E-Commerce?

• Geo: if you sell a finance product, a US-based company might expect QuickBooks integration, while an EU-based one might demand Cegid and GDPR alignment.

• Role: be crystal clear on who experiences the pain — and who makes the buying decision. Your tool might solve a daily headache for an SDR, but if the CRO is the one signing the check, your ICP is the CRO. Always anchor your ICP to the role that drives the purchase.

• Team size: different from company size, sometimes what matters most is the size and composition of the team you’re selling to. For example, if your product brings significant value specifically to finance teams with at least 4 people, define your ICP accordingly.

• Tools used: If your product fits neatly next to tools like Salesforce, HubSpot, or QuickBooks, or replaces a clunky part of them, bake that into your ICP. If 80% of your best-fit users already use Tool X, it’s a signal.

• Willingness to buy: don’t just describe who your ICP is — define what makes them ready. Ideal buyers have the pain now, the authority to make a decision, and budget to move.

  “Many founders aren’t ruthless enough when defining their ICP. It should feel almost absurdly narrow. That precision makes product decisions and sales conversations dramatically easier because your customers all share the same needs. Early on, prioritize ICPs that are easy to close and deeply engaged. Your first users should help shape the product with serious usage and real feedback.”

  – Vasco Alexandre, CEO and Cofounder of Dotfile

It’s time to test and refine your ICP

The purpose of working with design partners (DPs) is to see who gets the most value from it, and just as importantly, who doesn’t. In the next section we’ll dive into how to run a successful DP program.

“You start to define your ICP through three key signals: who you can engage in meaningful conversations, who you can convert, and who actually sticks with your product. To test that third dimension, we made a deliberate choice early on: avoid locking design partners (or later customers) into long-term contracts. Instead of offering annual commitments, we went with one-month engagements. This gave us the agility to iterate quickly and see who truly fit. While shorter contracts made deals easier to close, they also led to higher churn. But that churn was incredibly valuable as it clearly showed us who wasn’t part of our ICP, and helped us sharpen our focus moving forward.”

– Hugues Renou, CEO & Cofounder of Tengo

Once you’ve selected your DPs, it’s time to test your assumptions. If you’ve defined multiple ICPs, this phase will help you see which one engages most deeply with your MVP. Pay close attention to usage patterns, feedback, and the key metrics that you’re following - they’ll help you refine key ICP attributes like tooling fit, team size, or urgency of the problem.

But your ICP shouldn’t be treated as fixed after this phase. It’s a living hypothesis, and ongoing refinement is essential. Continue testing through outbound campaigns, inbound leads, and sales conversations. For Spendesk, they quickly realized that small companies weren’t the right ICP fit:

“When speaking with a three-person company, it was clear that managing spending wasn’t yet a significant issue; they would simply share one company card around the table, without any real friction. However, we observed that once a company reached a certain stage of growth — when it began structuring itself more formally — things changed: the finance manager or CFO typically became the one holding the company card, and that’s when the real pain surfaced.”

– Rodolphe Ardant, cofounder of Spendesk

Feedback loops between ICP and MVP

When you’re validating your early market, the instinct is often to tweak your ICP when things don’t click to look for a “better-fit” customer. But sometimes, the insight isn’t about changing who you’re targeting. It’s about changing what you’re offering.

As new customer segments start to emerge, you may realize they use your product differently and could unlock more long-term value. You might shift your ICP toward this new segment and evolve the product to better serve their needs. It becomes a flywheel: a more valuable product attracts better-fit customers, which further informs product direction.

“At first, our customers were very small teams but as we started signing slightly larger companies, with five to ten users, we noticed a critical pattern: they weren’t just using phones. They were using entire tool stacks to manage their operations. We quickly realized that if Aircall wanted to be truly valuable, it would need to develop integrations.

If we had stayed focused on very small teams, with a lightweight product, we might have built a great business, but probably not a product that needed deep integrations. So it wasn’t just product logic, it was also customer evolution that led us toward integrations and being part of a broader tool stack.”

– Olivier Pailhes, cofounder of Aircall

From 1 to 10 clients: Design Partner program

Why run a DP program?

In the earlier chapter on Product, we covered how crucial design partners are for validating your MVP and sensing early product–market fit. Now let’s look at how you can turn your design partners into your first 10 paying customers.The DP Program is your first real confrontation with the market. It’s where theory meets friction. You’re not just talking to potential users, you’re solving real problems for real people, inside real companies.

You’re using your time, your product, and your expertise to help them achieve something specific. When you do, you’ve not only validated the need, you’ve earned your first customers.

How to find your DPs

Here’s how to find your design partners:

• Tap back into discovery. The people you spoke with during discovery are a great start. If they fit your ICP and showed a real need, circle back.

• Leverage your network. Design partner programs are built on trust, and warm intros go a long way. These folks are more likely to give you a shot, and more forgiving of early bugs.

• Go public. Post a “call for design partners” on social media. Be specific: what you’re building, who it’s for, what you’re offering.

• Run targeted outbound. Send small batches of highly personalized messages that fit your ICP and cold call if necessary.

If you have interested prospects but your MVP isn’t ready yet, start nurturing them. What works well:

• Short, regular product update emails

• Behind-the-scenes content (screenshots, design decisions, early wins)

• Intimate invite-only events or roundtables to bring early believers together

“Our approach was completely outbound, primarily through discovery calls. We would tell potential partners about our Design Partner Program using a Deck and invite them to join.”

– Hugues Renou, CEO & Cofounder of Tengo

It shouldn’t be free

Yes, DPs should pay. A fixed fee, for early access over a defined period. At
the end of that period, they get a discount if they decide to become long-term customers. But the fee matters — if you can’t find five companies ready to pay for your product, even in its earliest form, you probably shouldn’t be building it.

“Have the ‘Willingness to Pay’ conversation as early as possible — ideally during the first user interviews... If they respond vaguely or seem unsure, that’s a red flag.”

– Jordane Giuly, Cofounder of Spendesk

It helps qualify serious partners and creates a clear expectation of value. The price doesn’t need to be high, but it should be high enough so that it’s still a commitment. For a bank, $1000 might feel like nothing and they’ll treat it that way. But for a startup or scaleup, that same amount could signal real buy-in. Find the number that makes your DPs care.

We often recommend splitting the payment, 50% upfront when the product is delivered, and 50% at the end of the program, once value has been proven.

“People only truly value what they pay for. If they don’t pay, they won’t use your product with the same intensity, and they won’t give you as relevant feedback. So it’s good to always try to monetize from the beginning.”

– Vasco Alexandre, CEO & Cofounder of Dotfile

“We monetized about two years in. 95% of people said they were willing to pay. But that also meant we clearly waited too long. Our pricing model was overly generous. In hindsight, it was a mistake not to be more firm: either you pay, or you don’t use it.”

– Christophe Pasquier, CEO & Cofounder of Slite

“Setting your price is a craft, not a formula. Mathilde Collin, cofounder of Front, famously asked early design partners to put a price on the value it created for them telling them, “It won’t be free, so tell us what it’s worth to you per user.” They could pay in money or even physical gifts. This approach grounded pricing in real perceived value. It’s not about getting the number right immediately, but about deeply understanding where the product sits in the customer’s mind.”

– Quentin Nickmans, Partner and Cofounder of Hexa

Think of it like a sales process - because it is

A Design Partner Program isn’t just about feedback, it’s your first sales motion. Yes, your MVP might still be rough, but you’re still selling it. And the fundamentals of sales apply.

You first call with potential DPs should be structured like this: 15 minutes of discovery to understand the prospect’s need 15 minutes on your DP deck — see structure below

Use the BANT framework to qualify:

• Budget: Do they have the budget or flexibility to pay?

• Authority: Are you speaking to the decision maker?

• Need: Is the problem clearly defined and urgent?

• Timing: Can they start using it within a reasonable timeframe?

If the lead isn’t a fit, it’s okay to cut the call short. Respect your time and theirs.

If they’re not the decision maker but show real interest, set up a second call with the right stakeholders.

Every call should end with a clear next step. After each call, define exactly where the lead stands:

• Not a fit → Close the loop and move on

• Not now → Add to your nurture list and check in later

• Qualified → Send an offer (not a pitch deck, an actual offer)

• Second call → Only if you need other teams (product, dev, ops) to assess feasibility or finalize with the decision maker

It’s perfectly okay to say no to a prospect. Don’t dilute your positioning by accepting every edge case.

“What really matters in a Design Partner Program isn’t how many partners you have, but how deeply they’re committed. Some of our most valuable partners couldn’t afford to pay but they were fully aligned with our vision and deeply invested in shaping the product. Charging can help filter out those who aren’t truly motivated, but it also risks excluding high-conviction partners who bring immense value. The key is to assess motivation, not just willingness to pay. In the early days, learning is far more valuable than cash.”

– Christophe Barre, CEO & Cofounder of Tandem

The design partner deck

At this stage, you don’t need a polished website or a lead qualification flow. What matters is having conversations via direct outreach or warm introductions. You won’t have a finished product, so what you’re really selling is your vision and your approach.

Use a simple, clear deck to explain what you’re doing and why it matters. Here’s a structure that works:

1. Your value proposition — one sentence, crystal clear

2. Market pain points — three key problems you’re solving

3. Feature overview — the MVP that addresses the pain points. If your product isn’t ready yet, show your wireframes.

4. Your team — showcase your expertise and credibility in their industry, highlighting how you can deliver value beyond just the product
   at this early stage

5. Design Partner Program — how it works, what they get, and what you expect

6. Next steps + contact info

Sending the offer

When sending an offer, one of the most important levers you have is urgency. Always include a validity date to encourage faster decision-making, it sets clear expectations and keeps momentum going.

It’s also a good practice to copy your CTO when you send the offer. They can follow up with next steps on implementation or onboarding. This adds credibility and gives the prospect a direct line to your operational team which builds trust.

Once the offer is accepted, send over your Design Partner agreement.

If your product isn’t ready yet, a Letter of Intent (LOI) is a solid alternative. It formalizes the commitment while giving you time to finish building.

Making the most of your DPs

Avoid stretching the program too long. Three months is a good rhythm, it gives you time to iterate without dragging things out.

Your goal during that time is to learn how your partners are actually using the product: what works, what doesn’t, and what they still need.

Set the right structure to encourage feedback and collaboration:

• Create a shared Slack channel: so they can reach you easily, ask questions, or flag bugs in real time.

• Set clear rituals: a weekly check-in call (to review issues, updates, new needs), and a monthly deep-dive workshop on a specific topic or use case.

When the program ends:

• Offer a discount or price lock to reward their early commitment and help convert them into long-term customers.

• With their consent, turn them into advocates — case studies, testimonials, webinars, or simply quotes on your landing page.

These partners are your first believers. If the program is well-structured, they’ll often become your best growth drivers.

“My goal was simple: that on a Friday night, when they’re out to dinner with friends, they’d proudly say, ‘I’m helping this small company that’s going to become huge.’ You want your design partners to feel like they’re part of the story. Invite them behind the scenes: bring them into your office, introduce them to your team, share your wins, your struggles, your fundraising journey. That shared sense of adventure is what they buy into. That’s how they become your strongest allies.

We ran weekly check-ins with operational teams and monthly reviews with decision-makers. But beyond the structure, trust is everything. I personally spent a full day with each partner, in their office, just observing how they worked, getting to know their team. You have to invest in the relationship and genuinely care about the problems they’re trying to solve.

These people aren’t just early users, they’re true partners in the project. And many of them became much more than that. Some even turned into angel investors.”

– Hugues Renou, CEO & Cofounder of Tengo

From 10 to 100 clients

GTM should be founder-led, still

Just like design partner sales, getting to 100 clients should be founder-led.

Because at this stage, every sales conversation is an opportunity to learn, not just to sell. You’re not simply pushing a product; you’re refining messaging, identifying objections, validating your ICP, and adapting your pitch in real time. No one is better positioned to do that than the founder.

In fact, founders should aim to personally lead sales until at least $1M in ARR, and ideally close the first $50K themselves.

“In the early stages, prospects are often buying as much into the foun- der as they are into the product. They’re investing in the team and the potential they see in you. They may tell themselves, “The product isn’t perfect today, but I like the approach and I trust this team will get there if I support them and give feedback.”

– Vasco Alexandre, CEO and Cofounder of Dotfile

Leverage your DPs with a lookalike strategy

Reaching this stage is a meaningful milestone, it signals that you’re starting to understand who your ICP really is. Now it’s time to put that insight to work and validate it at a broader scale.

At first glance, outbound might seem simple: build a list, enrich the contacts, send sequences, wait for replies. But in reality, that spray-and-pray approach almost never delivers real results - especially in early-stage B2B.

To break through, you need to move toward what we call “not-so-cold” outbound: outreach that feels warm from the first touch. One of the best ways to do this early on is by running a lookalike strategy, using your existing Design Partners as your foundation.

Start with your refined ICP, shaped by what you’ve learned from your current Design Partners. Look closely at the patterns: What industries are they in? What’s their company size and where are they located? Which roles are involved in using and championing your product? And how, exactly, are they using it day to day?

From there, identify lookalikes. Who are their direct competitors? Which companies serve the same customer segments? Who operates in a similar space, with comparable team structures, workflows, or challenges? These are your warmest leads, because they’re most likely to understand the value your product delivers.

When you reach out, keep the message simple. A line like:

“We’re the solution currently used by [Company] to manage [problem]...”

It immediately adds relevance and builds trust. It frames your product as pro- ven and significantly boosts reply rates.

The magic of this strategy is that it compounds. Every successful Design Partner opens up a network of similar companies to explore. If done well, outbound doesn’t feel like cold outreach anymore, it becomes a warm expan- sion of a trusted ecosystem.

To keep the momentum going, focus your efforts on clients who’ve seen the most value from your product. These become your reference points not just for outbound, but for use cases, testimonials, and messaging.

“Once we had those early users, something powerful happened: they started recommending us... That allowed us to expand into a second circle of users... With a clearer understanding of our ICP and the key verticals where Spendesk could bring value, we launched an outbound strategy.”

– Rodolphe Ardant, Cofounder of Spendesk

Not-so-cold outbound

Another way to achieve that warm intro and to avoid sounding generic, is to start by identifying subtle cues that show real intent or context.

Start by identifying signals of intent

Before you write anything, gather insights that show what matters to your prospect right now.

Look for:

• Social activity: Recent social media posts, comments, or who they follow — all can hint at current interests or priorities.

• Public content: Job postings, blog articles, press releases, or podcast inter- views often reveal strategic initiatives and upcoming needs.

• Community engagement: Notice which Slack groups, Discords, or events they’re active in — it shows how and where they like to engage.

• Potential value-aligned offers: Can you give them something useful for free? A tailored teardown, a template, a short consult, or a playbook that matches their current goals.

Reach out through the right channels

Email is just one option and not always the best one. Early-stage outreach can benefit from meeting people where they already are. Try:

• LinkedIn DMs (especially if they’re active there)

• GitHub (if you’re targeting developers)

• Slack or Discord communities

• Events or webinars they attend

Setting up a CRM

In the early days, a complex CRM isn’t necessary. If you haven’t used one before or aren’t comfortable with setup, it’s better to wait until you make your first business hire and use a no-code tool and flexible tool for now.

At this stage, the goal isn’t complex tracking, it’s clarity. You need to know exactly what’s happening in every deal, at every step of the process. That means building a pipeline with zero black hole where every stage reflects a clear action taken.

If you’re seeing deals pile up in one stage (especially post-initial contact), it’s usually a signal that something’s broken: unclear next step, misalignment, or low engagement.

Here’s a simple CRM flow we’ve seen work well at the earliest stage:

• Inbound → Lead requested a demo or reached out

• Qualified to contact → You’ve reviewed the lead and it fits your ICP

• Outbound to start → Lead identified or imported, ready for outreach

• Contacted → You’ve sent a message or launched a campaign

• Meeting planned → A call or demo is scheduled

• Offer sent → A formal proposal is out

• Won → Deal closed: paying customer or signed design partner

• Come back later → Good lead, not the right timing — worth nurturing

• KO / not interested → Clear no, don’t include in product updates or events

• Not relevant → Bad fit (wrong company, role, etc.)

Always set up a clear follow-up, and never end a call without establishing the next step or sending an offer.

“Sales is not an art form that needs to be reinvented for every company. There are tried-and-tested methods for selling software that have been working for years. Entrepreneurs have plenty of space to innovate but innovating on sales methodology is usually not a good idea. Instead, relying on proven sales frameworks can be incredibly valuable. Once we started applying real methods and processes, and the whole team began operating in the same structured way, it became much easier to create repeatability in our sales motion.”

– Vasco Alexandre, CEO and Cofounder of Dotfile

Don’t limit content to your product

A common early-stage mistake is making your content too narrowly focused on your product.

Say you’re building a company meal voucher platform. You might be tempted to stick with calculators or pricing comparisons. But that’s limiting.

Your product lives in a broader narrative — employee happiness, workplace culture, team rituals, HR benefits. That’s where the content opportunity really lies.

The best editorial strategies are anchored in a larger ecosystem. They give you dozens of angles to work with and build long-term relevance.

Front nailed this by acquiring its first hundred customers through content pieces: instead of focusing solely on the product, founder Mathilde Collin talked about what it was like to build a startup, how to run 1:1s, transparency at work, what it’s like to raise a seed.

“Front gained its first hundred customers through content pieces that I wrote. The founders who can really stand out with content are the ones who are both confident in their point of view and genuinely enjoy writing. Personally, I loved writing. And I didn’t even write directly about Front that much. I wrote more about the founder journey and things I felt were missing or not spoken about transparently.”

– Mathilde Collin, Founder of Front

Understand where your audience hangs out

Map out the marketing landscape: who the legacy players are, which challengers are on the rise, how the top companies position themselves, and what kind of content they’re putting out into the world. Study everything: the tone they use, the structure of their blogs, the angles they take on social media, the formats that resonate most, and the keywords that dominate the conversation.

Don’t stop at the obvious channels. Go wide. Scroll through Product Hunt, Reddit, GitHub, Hacker News, and LinkedIn. Everything is signal. The goal is full saturation — to be so deep in the space that subtle patterns start to emerge without you even looking for them.

Once you’ve absorbed the landscape, patterns start to emerge. You’ll begin
to recognize which topics dominate the space and which content formats or lead magnets consistently gain traction. You’ll see which marketing actions are actually visible, whether it’s polished inbound campaigns, memorable branding moves, or smart outbound strategies. You’ll also notice key trigger moments in the calendar: big conferences, regulatory shifts, product launches, or budget cycles that spark urgency in your target audience.

And equally important, you’ll understand where your audience actually hangs out — whether it’s LinkedIn, GitHub, Reddit, or somewhere more niche.

With all that insight, you can shape a clear point of view for your personal and company brand, identify 3–4 immediate marketing plays worth running, and begin sketching the contours of an editorial universe that your brand can own.

The power of mini-products

Mini-products are one of the smartest ways to generate inbound interest in your product by providing value to users.

They can take many forms: a free template gallery, a stripped-down version of your product, a niche toolkit tailored to a specific use case, or even a curated database. What they all have in common is that they solve a real, immediate problem for your target audience - without requiring a full product signup or sales call.

The best mini-products are built from your core product.

It’s about turning a valuable slice of your product into a marketing asset - something people can land on, use instantly, and get value from. Rather than explaining what your product does, you let them experience it.

You need to think carefully about the user flow to drive conversions to the real product: should users sign into the product to access the mini-product? Where should you place CTAs for maximum impact?

Why it works:

• Low friction, high value. Mini-products don’t ask much from users—no demo, no long onboarding. Just instant usefulness.

• Distribution-friendly. They’re easy to share, link to, or embed in other content—making them perfect for SEO, community, and virality.

• Audience-aligned. When done right, they address a precise need your ideal customer already has.

• Brand-enhancing. They position you as helpful, thoughtful, and an expert in your space.

Let’s look at some notable examples of mini-products launched by Hexa companies:

The myth of PLG

Product-led growth (PLG) is a powerful strategy but it’s far from the only one. A common myth is that PLG means the product “sells itself.” That’s not just inaccurate, it’s dangerous. It gives founders a false sense of security, leading them to avoid building a real GTM motion. And while skipping GTM might feel easier in the short term, it’s a risky bet in the long run.

In reality, PLG and sales are not opposing forces, they are deeply complementary. A successful PLG strategy often creates momentum, but that momentum needs to be captured, guided, and amplified. Even with a viral product, your team may still need to invest heavily in outbound efforts, customer success, and strategic sales to truly scale.

“What does it even mean to be ‘Product-Led’? Too often, founders take it to mean ‘not Sales-Led’ — as if building is somehow more virtuous than selling. But that mindset lets them hide in their comfort zone. They say, ‘We’re product-led, great products sell themselves,’ and then avoid sales altogether. They demonize outbound. They try to solve every problem with more product. And in doing so, they run away from product–market fit. We should stop trying to be PLG. PLG isn’t an identity, it’s a go-to- market motion. There’s no such thing as a ‘PLG company,’ only ‘PLG strategies’.”

– Mehdi Boudoukhane, CEO & Cofounder of Cycle

“We were caught up in the early wave of product-led growth and tried
to mirror Slack’s model. But we were not Slack. The usage behavior and economic units were completely different. With Slite, users come in with a clear goal, set up an internal knowledge base and improve collaboration. If they see value, they pay. If not, they leave.”
– Christophe Pasquier, CEO & Cofounder of Slite

Upflow had a similar experience. They launched a free, lightweight version of Upflow, but they quickly realized that it wasn’t going to sell itself. They realized it was more about product assisted sales than product led growth.

“At first, we imagined it would work like a textbook product-led strategy, where people would discover the tool, sign up naturally, and move smoothly through a conversion funnel, but in reality, it didn’t happen that way, it still required a lot of human interaction - our salespeople calling up doing demos, and personally guiding users through the experience.”

– Alex Louisy, CEO & cofounder of Upflow

Customers without borders

One of the biggest advantages of building software is its global reach - you’re not limited by geography. As long as your product complies with local regulations and addresses real user needs, it can be sold almost anywhere. Yet many European startups limit themselves to local markets far too early.

In the early-stages, you of course won’t open up an international office - but here’s what you can do to sell globally:

Hire native-speaking talent locally

Recruit sales or support reps from key target markets, even if they work from your HQ or remotely. You’ll gain native fluency and market insight without needing a local entity. It’s also a strong move for diversity and inclusion, which we’ll explore later in the ESG section.

“One thing I really recommend is hiring English-speaking locals wherever you’re headquartered before trying to fully set up in the U.S. In the early stages, when you’re only hiring a small number of people, finding a native English speaker locally is much better than having French teams cold-call American prospects with heavy accents and missing cultural references.”

– Alex Louisy, CEO & cofounder of Upflow

Follow your growth

Don’t wait until everything is perfectly in place to expand internationally, follow the momentum. If you’re seeing unexpected inbound interest from a particular country, lean into it.

“Our logic was simple: wherever we had inbound, we did outbound. Our mindset was: why limit ourselves? We’re on the internet, there are no borders... If someone wants your product, why say no just because of geography?”

– Jonathan Anguelov, Cofounder of Aircall

“With a product like CRM there’s no real geographic limitation, which opens up a huge international opportunity. So we built for that from the start. We also naturally developed a strong U.S. focus early on which was largely driven by our Product Hunt launches, where the majority of users discovering us were based in the U.S. Today, the U.S. represents about 40% of our revenue, which confirms the strength of that strategy.

– Simo Lemhandez, CEO and Cofounder of Folk

Get business angels onboard so they can open up markets and connect you to the right people

You’ll find more on this in our discussion with Evan Testa, CEO and cofounder of Roundtable, right after the fundraising chapter. But here’s a sneak peek at the strategy:

“Before entering a new market, I’d go find the right people one year ahead, usually people I knew I wanted on my cap table, and incentivize them early. In Germany, for instance, I connected with key folks, told them, “I’m not launching here yet, but I will,” and brought them in as early believers. By the time we officially launched, they already understood the value, the mission and they became our gateway to that market. Whether they were clients or connectors, they helped with hiring, distribution, and trust-building. It’s local distribution, built-in.”

GTM Rituals

By now, you’ve probably caught on to a recurring theme in this book: rituals are key. To make meaningful progress, your GTM efforts need both structure and cadence. That means establishing two core rituals from the very beginning, quarterly for strategic planning, and weekly for execution.

Quarterly GTM rituals

These are your high-level planning sessions, the moments to define clear goals, align with the broader company mission, and decide where to focus your energy in the upcoming quarter.

OKR guidelines:

• Set clear, quantitative OKRs - especially for sales. What exactly do you want to achieve this quarter? Put numbers to it.

• Balance two priorities: growing your current accounts and generating new ones. Don’t over-index on one at the expense of the other.

• Align your GTM OKRs with your company’s core mission, and more often than not, with your product roadmap. In early-stage teams, GTM often follows product OKRs.

• Break down your Objectives into Key Results that are tangible, testable projects. These are your experiments.

Experiment guidelines:

• Prioritize experiments that can be built, launched, and measured in under four weeks.

• Aim for a mix of inbound and outbound tactics to diversify your approach. Be precise.

• Rather than “launch SEO,” try: “Publish a 10-article SEO series focused on construction tenders targeting procurement leads.”

• Always justify the experiment: Why is this a priority? Why this audience? What signal are you betting on?

• Use a combination of market data and informed intuition. In the early stage, gut instinct often plays a valid role.

• Encourage creativity. Early GTM is about testing boundaries, not staying inside them.

Example of quarterly experiments:

Weekly GTM sprints

Once your quarterly plan is in place, it’s time to shift into execution mode. Each week, you’ll break your experiments into manageable tasks and move them forward.

The goal here is consistency: small, focused actions that compound over time.

Every week, ask:

• What are we shipping?

• What are we learning?

• Are we moving closer to validating or invalidating the experiment?

• Do we scale this or do we move on?

Example of a weekly GTM sprint table:

To get the Hexa Roundup delivered every month to your inbox, don’t forget to subscribe below - or go to your account settings to turn of Hexa Roundup emails.

Subscribe now

📖 The 10x Method: New chapters

Four new chapters of Hexa’s sold-out book are now live, one dropping every week.

II. Finding & validating the idea — A startup idea is a framework: problem, solution, and GTM strategy. This chapter breaks down the three categories of startup ideas, why every great startup has to innovate on at least one dimension, and the validation process Hexa uses before writing a single line of code — including our ideas that didn’t make the cut and exactly why. → Read the chapter

III. Choosing your cofounder — After matching 100+ founders, we’ve identified the three ingredients that make or break a cofounder relationship: complementary skills, alignment on five non-negotiables, and radical transparency from day one. The chapter covers how to run the search like a sales funnel, what questions to ask, and why you should bring in a coach before things go wrong, not after. → Read the chapter

IV. Building your MVP — The biggest mistake founders make isn’t building the wrong thing — it’s building too much, too soon. This chapter covers how to scope ruthlessly, why your designs should be super basic, and how to work with design partners to get real feedback instead of theoretical reactions. → Read the chapter

V. Hiring early talent — Founders should spend 30-40% of their time on recruiting in the early days. This chapter covers who to hire first (engineers, then a founder associate — not a PM), how to run scorecards from day one, structure equity, and know when — and how — to part ways. → Read the chapter

🤏 How going narrow landed Yoobic €100K contracts

It can be tempting to cast a wide net to find customers. Fabrice did the opposite, narrowing until Yoobic was the obvious choice for fashion retail chains managing deskless workers across hundreds of stores. Hexa Partner Mat Vaxelaire sits down with the Yoobic cofounder in the latest episode of PMF Collector to break down the methodology behind it: validating only through paying customers, proving ROI before asking for budget, and why almost nothing from the European playbook transferred when they expanded to the US.

→ Watch the episode

🔧 How to scale from basic RAG to 200+ e-commerce sites

Louis Pinsard built Dialog, an AI shopping assistant now live across 200+ ecommerce stores, starting with a LangChain tutorial in production. Scaling meant solving problems no tutorial covers: catalogs that break your retrieval, queries too vague to search with, and six months of iterating by gut feel before building a proper evaluation system. In this Build with AI episode, Hexa Product Lead Pierre sits down with the CTO to go through every layer of the evolution and what he’d do differently from day one.

→ Watch the episode

Portfolio news

🏥 Parallel (HX24) secures $20M to make hospitals more efficient with AI

Paul Lafforgue, Chris Rydahl and the Parallel team just closed a $20M Series A led by Index Ventures, with Frst, Y Combinator, and angels. Their AI agents already live across several dozen public and private hospitals in France, automating the administrative workflows that eat up to 30% of healthcare spending. What makes Parallel’s approach distinctive is speed: their agents run directly on top of existing hospital software, no lengthy integrations, live in as little as a week. Very proud of the team! Learn more.

🔐 Hodor joins Hexa Sprint to secure AI agents

Karim Boussetta and Timothé Husson Bromberg founded Hodor after spotting a gap that’s only getting bigger: companies are plugging AI agents into everything, but nobody is controlling what they’re actually allowed to do. A misconfigured agent can give 90% discounts to everyone, send seven follow-ups per hour, or expose millions of records before anyone notices. Hodor is the identity and policy layer that prevents exactly that, joining Hexa Sprint’s AI for Work vertical led by Thibaud Elzière. Learn more.

Hexa news

🔧 Tanguy Goretti joins Hexa as our first ever CTO

Tanguy is a repeat entrepreneur who has been building for 20+ years, best known for co-founding Cowboy, raising €100M+ and building a fully integrated hardware and software platform connecting 80,000+ riders across Europe. He joins to work hand in hand with founders on the most technically ambitious AI companies in Europe - as AI shifts from product feature to core company architecture, the decisions founders face are increasingly deeply technical. Learn more.

🗣️ Ship your voice agent at Hexa next week

Next Thursday, we are co-hosting an evening with HappyRobot, Early Builders and OpenAI dedicated to voice and multi-modal AI agents. Come with a teammate and a problem in mind - mentors from OpenAI and HappyRobot will be in the room all night to help you build on HappyRobot, the infra platform behind some of the largest enterprise AI deployments right now. Start at 6pm, demo at 9pm. Spots are limited, apply here.

Jobs

• Founder Associate Intern at Hexa

• Founding Full-stack Engineer at Waniwani (HX26)

• Founding Full-stack Engineer at Verso (HX26)

• Founding Engineer at Panora (HX25)

• Founding Engineer at Plato (HX26)

• Legal Product Manager at Plato (HX26)

• Talent Acquisition Partner at Hexa

• VC Analyst Intern at Hexa

Thanks for reading! Follow us on Linkedin and Instagram to stay up to date between issues.

Cheers,

The Hexa Team

Founders often assume that structured hiring and people practices, like using scorecards in interviews, running regular 1:1s, or giving formal feedback, can wait. The belief is that those things are for later, once you’ve scaled. The truth is, hiring and retaining talent at the earliest stage isn’t fundamentally different from doing it at later stages. But it matters more. When you have a team of one or two, every person carries outsized weight. Hiring the right people, and keeping them, requires just as much structure as it does at 50 or 500 people.

This is one of the most critical chapters in the book. Early hires shape everything. We’ve seen it again and again: get them right, and you build a strong foundation. Get them wrong, and the risks are massive. Without the right early team, your product will suffer. You’ll struggle to acquire customers. Your brand won’t take shape. Your metrics will lag. Everything downstream depends on getting this part right.

The first people are setting the standard. At Spendesk, it was early interns who coined the phrase “ship, ship, ship,” capturing the company’s bias for action and fast execution. When those early hires moved on years later, some of that unique energy left with them.

At Swan, the first 20 hires are still with the company. They carry deep institutional knowledge and continue to reinforce the culture that made the business successful in the first place.

Early employees hire the next generation. That’s why the bar needs to be high from the very beginning.

In this chapter, we’ll show you how to find, attract, and retain the right early hire. From identifying who to bring on board first, to running a great hiring process, to building the kind of team culture that keeps top talent engaged and growing with you.

Who to hire

Time well spent

As a founder, hiring is one of the most critical responsibilities you will take on. It is not something you can delegate in the early days and so understanding the process firsthand is essential. You should be prepared to spend 30–40% of your time on recruiting and HR-related tasks in the early stages.

“Consider the first 10 people you hire with the same level of rigor as you would for a cofounder. Some first employees can eventually become late- stage cofounders, their impact can be just as crucial.”

– Quentin Nickmans, Cofounder and Partner at Hexa

“A company isn’t built by founders alone, it’s built by the entire team. But it can be undone by the founders alone... That’s why your early hires are so important: they set the tone for the culture, the ambition, and the way your company thinks as it scales.”

– Evan Testa, CEO and cofounder of Roundtable

The only way to learn is by doing. Before hiring someone to handle a function, you need to immerse yourself in it. This applies to coding, sales, product, and most importantly, recruiting. Sourcing, reaching out to candidates, and running interviews should be part of your initial efforts. Sending LinkedIn messages, following up, and refining your pitch will teach you what works and what doesn’t. Just like hiring a salesperson without having spoken to customers yourself makes little sense, hiring a recruiter too early disconnects you from a fundamental process.

That said, you don’t have to do everything entirely alone. It introduces too many biases and you end up making poorer decisions. While hiring a full-time recruiter as one of your first employees is unnecessary, you can leverage external support. Engaging a freelancer, a recruitment agency, or an RPO (recruitment process outsourcing) consultant can help you source and generate leads. However, this only works if you already understand the basics of hiring, otherwise, you risk relying on a system you don’t control.

Engineers

Your first 1 to 3 hires should be engineers, ideally full-stack. What’s most important is flexibility: even if they’re primarily backend or frontend, they must be willing to work across the stack. In the early days, your product will evolve quickly, and you’ll need people ready to adapt. Hiring someone too specialized, and unwilling to step outside their core expertise, can box you in before you’ve even figured out exactly what you’re building.

As discussed in the chapter on cofounders, the technical founder must be hands-on, not just managing, but actively coding. During those early months, this is non-negotiable. At this stage, you’re not building a tech organization, you’re building the product.

A frequent challenge we’ve seen? Founders moving quickly to hire freelancers. It can seem like the fastest path forward, you have a clear idea of what you want to build, and you’re eager to ship. But that early speed can lead to setbacks.

It’s important not to outsource your first build. This is the most critical phase of product development, and it needs to be done by people who are committed to the long-term success of the company. Freelancers won’t lay the groundwork for scale. They won’t set up clean architecture, strong processes, or thoughtful systems.

The only time we recommend working with freelancers is for very specific, tightly scoped tasks, for example, coding a specific part of the product. But you should never rely on freelancers to lead or drive major projects during this phase. And they should never replace a hire you will eventually make.

Coding is also strategy. And that’s why the technical founder should be in the trenches, working side-by-side with the first hires. That’s how you build a strong foundation, and bake long-term vision into the code from day one.

Yes, hiring great engineers takes time. But not as much time as you think. Especially if you follow the steps outlined in this chapter.

“One thing that’s been key is involving engineers deeply in the product. Product scopes and defines the goals, but engineers co-design the solution with them. That tight collaboration has been part of our DNA from day one. It’s also what helps attract truly valuable founding engineers: the kind of builders who don’t just want to execute, but want to be part of the creation phase, shaping the product from the ground up.”

– Titouan Benoit, CTO and Cofounder of Dotfile

The founder associate

Your first business hire should be a founder associate, a highly versatile, execution-oriented generalist who can plug into whatever the company needs most. In the early days, that could mean supporting customer calls, launching growth experiments, or building out investor materials. Think of this person as your right hand.

Range matters more than specialization at this stage. You need someone with broad skills, and even more importantly, the ability to quickly adapt, because priorities and scope will inevitably shift as the business evolves. Over time, one functional area like growth, account management, customer success, or marketing will naturally start to take precedence based on the company’s needs and their individual strengths.

This role becomes even more critical during fundraising. When founders are out meeting investors, someone has to keep the machine running. A great founder associate can step in as a chief of staff, prepping materials, managing the process, and ensuring continuity so nothing breaks while the founders are focused elsewhere.

Founders often underestimate just how stretched they’ll be once the company starts gaining momentum. A founder associate helps extend your capacity. They’re not there to “own” departments, but to push forward projects, uncover problems early, and make sure critical tasks don’t fall through the cracks.

“Once we had landed two or three clients, we realized we needed someone to help drive and support the business side more consistently. So, we hired a highly versatile profile, someone capable of handling a bit of everything: outbound sales, marketing content creation, direct sales conversations, and even product demos when needed.

For a long time, we operated with this highly versatile profile, because at that stage, the business was still fundamentally founder-led. It was only much later, once we had matured as a company, that we started hiring more specialized sales roles.There’s no definitive hand-off where a founder can fully step away from sales, it simply doesn’t happen. Even today, while the team handles certain segments successfully, I’m still actively selling: whether it’s entering new markets, expanding into new geographies, or building partnerships. As the company grows, sales remains a constant part of the founder’s role, just in different forms.”

– Vasco Alexandre, CEO and Cofounder of Dotfile

Don’t hire a product manager early on

In the early days, the product vision should live squarely with the founders. Great product managers naturally want to own the product roadmap and vision. If you hire one too early, they may start to behave like late-stage co-founders, leading to misalignment and friction. At this stage, that would be a mistake.

An alternative approach? If delivery is starting to slip or you need help with coordination, bring on a part-time freelance product owner. Just make sure the strategic direction stays with the founding team, at least until you’ve hit product-market fit.

You should hire a product manager only when your product is showing real traction and scaling execution becomes difficult. By then, founders will often be overwhelmed with new priorities like operations, fundraising, and hiring. That’s when bringing on a PM makes sense: someone who can formalize processes, take ownership of execution, and keep product development moving forward.

“In B2B software startups, your first product manager should come as late as possible — ideally after you’ve found product–market fit, not before. Until then, the founder is the PM. PMF isn’t something you can outsource; it’s something you earn. It demands obsession: talking to customers, shipping fast, watching how people react, and adjusting constantly. Bringing in a PM too early creates distance right when you need to be closest to the problem. It slows down the rapid learning loops that are essential at this stage. You don’t hire a PM because you’re unsure what to build — you hire one when you can no longer keep up with the demand.”

– Mehdi Boudoukhane, CEO and Cofounder of Cycle

Stay flexible

Some startups will need to make specialized hires earlier than others. The specifics will depend on your business.

For example, Roundtable, where legal complexity is central to the product, hired a Head of Legal very early, along with operations staff as part of the first few employees. It was the only way to get the business off the ground.

The key is to know what’s truly essential for your company to move forward, and hire accordingly.

Hire for attitude AND for skills

You’ve probably heard the phrase: “Hire for attitude, train for skills.” And it holds true, but more in the later-stage of company building. In the early days, that only gets you halfway there.

For your first hires, you need attitude and skills. You simply won’t have the time or management layers to train people from scratch. Early team members need to hit the ground running. They should already know how to operate in fast-moving environments, solve problems independently, and contribute from day one.

What you need to look for:

• Highly adaptable and curious. They’ll be wearing a lot of hats.

• Deeply autonomous. They don’t need hand-holding or constant direction.

• Entrepreneurial. Ideally, they’ve worked at a startup before or built something.

A common instinct when building your early team is to hire for your weaknesses, to fill the gaps. While that can make sense later, it’s not always the right first move. In the early days, what matters most is speed and execution.

“One of the least intuitive but most powerful moves you can make as a founder is to hire someone who shares your superpower. If you’re great at product, hire another strong product builder. It might feel counterintuitive, you’d think you should fill your gaps, but starting by doubling down on your strength makes delegation easy, hiring fast, and execution smooth. You’ll be able to transfer knowledge quickly, and free up your time without compromising quality. That time, you can then reinvest into tackling other areas, like sales or hiring, one by one. Only a founder has the grit to step out of their comfort zone and still give it everything they’ve got. By doing it themselves, they gain mastery over every aspect of the company. That’s how you grow your company: by reinforcing what you do best, then expanding deliberately.”

— Quentin Nickmans, Partner and Cofounder of Hexa

How to hire

Active search

There’s no way around it: if you want to find great candidates, you can’t sit back and wait for them to show up, especially when your company is still young and unknown. You have to go on the hunt. You need to put in the work, scour LinkedIn, and actively convince exceptional people to join you.

Define a clear search strategy: Think about where to find the right people. If you need a frontend engineer, target companies known for exceptional UI design.

Try the “indirect ask” trick: We’ve seen this work time and time again. Instead of asking someone directly if they’re interested, ask: “Do you know anyone great for this role?” This lowers their guard, and sometimes sparks interest in the role themselves. And if someone you’ve asked directly says no to the role, ask them if they know someone who could be great for it.

Leverage your network: Your best hires may come from referrals, but it’s not enough to send one message and hope for a response. Always follow up at least twice. Many great candidates don’t respond immediately, and persis- tence is key.

Track your response rates: If fewer than 20–30% reply to your outreach, something’s off. Are you targeting people who are too senior? Is your message clear enough? It’s important to test different messages and people to refine and improve your response rate.

Outbound is a numbers game, but quality still matters: Don’t just spam LinkedIn. Tailor each message. Reference their work. Make it obvious you’re reaching out to them, not just “someone.

“What’s interesting is that even though we didn’t hire a lot of people directly through outbound, it still paid off. Most people we reached out to didn’t join, but they referred someone in their network. Outbound actually created a kind of network effect. When you start reaching out, people often respond with, ‘Hey, I’m not looking right now, but I know someone who is.’”That ripple effect really worked for us.”
– Evan Testa, CEO and Cofounder of Roundtable

Build a brand early-on

Inbound recruiting for a young startup is a long-term game, but one worth starting early because these efforts will compound faster than you think.

• Start employer branding early: Even as a small startup, having a basic Notion career page with company values, images, and insights can make a big difference.

• Leverage LinkedIn: Regular posts about your journey, team milestones, and open roles help create visibility and momentum. Top talent often follows founders, not companies.

• Participate in events: Meetups, panels, and industry events are underrated for hiring. They build trust and help you connect with people who aren’t actively job-hunting, but might be open to the right opportunity. But be very selective about the ones you attend as they can become a significant timesink if you’re not too careful.

“Even when I wasn’t actively hiring, I made a habit of adding people on LinkedIn who looked like strong future candidates. It turned out to be incredibly valuable. When I eventually started recruiting for a PM role, many candidates already knew about Tengo from my posts. It made a real difference, the response rate was much higher.”

– Hugues Renou, CEO and Cofounder of Tengo

“I started posting on LinkedIn mainly for recruitment purposes. Finding great people, engineers especially, who share your mindset is very hard. By sharing thoughts, experiences, and opinions publicly, you naturally attract people who resonate with your way of thinking... It’s a kind of natural selection: people either connect with your energy or they don’t, and that’s exactly what you want in a startup team.”

– Paul Vidal, CTO and Cofounder of Collective

The job description

Most startup job descriptions are either too vague or too corporate. They either try to sound “cool” and end up confusing candidates, or they copy language from big companies that completely misses the early-stage reality. Neither works.

Your job description isn’t just a formality. It should excite the right candidates and filter out the wrong ones. Be clear about the mission, honest about the challenges, and specific about what success looks like. Make sure it also highlights the level of ownership early hires will have, because that’s what great early-stage candidates are looking for: real impact. Help them see it.

1. Title

• Choose a job title that people actually search for, not just one that sounds attractive. For example, if you’re hiring a “Founding Engineer,” that’s fine as a title, but make sure to include terms like “Software Engineer” throughout your job description so candidates can find the role when searching those common keywords.

2. About the company

• Remember, your company is new and unknown, so this section is crucial. Show what makes you unique and innovative.

• Early hires join startups because they care about the “why.” Start with a punchy paragraph that explains what you’re building and why it matters.

• It’s also a good place to emphasize how the first employees will play a huge role in shaping the company’s culture and impact.

3. What you’ll do (missions)

• Forget the laundry list of responsibilities. Instead, describe the core problems they’ll own. Early hires don’t need every bullet point spelled out but they need to know what’s broken and what they’ll be trusted to fix.
  → “You’ll be responsible for developing and maintaining scalable front-end and back-end solutions.”

• Early employees’ job descriptions change rapidly. The role can be completely different six months after hiring compared to when they first start. When writing the job description, think ahead and clearly communicate how the responsibilities will evolve from day one through the first six months.
  → “This role starts as an individual contributor, but within 6–9 months, we expect you to help hire and mentor additional team members.”

• Adding a section “What success looks like in 6 months” is a great way to be forward looking

4. Technical stack (for certain roles)

• List the technical stack for this role. This is common for engineering positions but should also be included for GTM, marketing, and operations roles. If you don’t have a technical stack defined yet and it will be the employee’s responsibility to create it, clearly state this in the description.

• Be sure to include a line like, “Don’t hesitate to apply even if you’re not familiar with every tool in our stack” to encourage more diverse candidates to apply.

5. Who you are (requirements)

Founders often make the mistake of listing too many requirements compared to the mission section, which makes the job description look unbalanced. This happens because founders typically have a clearer idea of requirements - that’s the easy part.

• Highlight just 3–5 traits or experiences that really matter

• Clarify what’s nice-to-have vs. essential

• The longer your requirements list, the less women will apply - so keep it short and impactful.

6. Why join (your company name)

• Here’s where you list your company’s values and benefits. You might not have defined your values on paper yet, and that’s okay. No need to rush.

• But you certainly have preferred ways of working. Are you async-first or meeting-based? Do you have monthly team lunches? Are there any special routines worth mentioning? This is your chance to let your culture shine, even if it’s still taking shape.

• Beyond the classic perks like health insurance and company snacks, early-stage companies offer unique benefits where employees can make a real impact. Make sure to highlight these opportunities here.

7. Hiring timeline

It’s crucial to outline the hiring process steps clearly so candidates know exactly what to expect. At the early stage, you likely won’t have many team members, or any at all, so focus on creating a quick process with key stakeholders. Even as a solo founder or duo of founders, it’s essential for candidates to meet several people. If you don’t have a team yet, bring others into the process, this could be peers, fellow founders, investors, or freelancers helping with recruitment.

1. Initial call – A quick introduction with the CEO/other person

2. Technical Interview – A discussion with the CTO about your experience and skills.

3. Meet the team – A casual conversation with a current team member to ensure a cultural fit.

4. Reference Calls – Final step before the offer. We conduct at least two reference calls, one with a former colleague and one with a former manager. Additionally, we offer you the opportunity to speak with one or two people who have previously collaborated with us, so you can ask any questions you may have.

The first interview is about convincing

Founders often treat the first interview like it’s about evaluation. It’s not. The first call is about attraction.

Your job is to make the candidate want to join, especially with outbound candidates. But also with inbound ones. They might not be sold on your company as it’s still largely unknown and so you have to hook them.

At the end of an outbound call, flip the script: “Do you want to be part of this process?”

Let them opt in. If they’re in, great - now you switch into interview mode.

From there on, it’s about assessment. Use a scorecard. Define what “great” looks like. But never forget: if you don’t win them on the first call, you may never get the chance to assess them at all.

“The goal of the first conversation is not to grill the candidate with technical questions. It’s to make them leave thinking, “I really want to work with these people.” As a new startup with little name recognition, you must actively inspire candidates. It’s a selling call. Another tactic I used: after a great first call, I would immediately introduce the candidate to another co-founder for a second “selling” conversation. It wasn’t announced in advance, but it made candidates feel valued and gave them the impression that the company was moving fast and cared about them.”

– Paul Vidal, CTO and Cofounder of Collective

The importance of follow-ups

Throughout the process, maintain a strong connection with the candidate. Quick check-in calls between each interview stage are key. These calls, or even a simple WhatsApp loop, help you:

• Keep the relationship warm - it helps show you care

• Track whether they’re interviewing elsewhere - you don’t want to lose the candidate because they got an earlier offer somewhere else

• Continuously gauge their motivation - it helps you know if they’re still en- gaged and interested in your company

Scorecards even at the start

Most founders aren’t experienced recruiters, and that’s completely normal. But it also means you’re especially vulnerable to bias. You might overvalue charisma, a shared background, or someone who just “feels right” in the moment. That’s why scorecards are so valuable at the early stage. It is a structured framework that helps you evaluate candidates based on clearly defined criteria, not just gut feeling.

A good scorecard also makes rejection easier. Instead of debating feelings, you have objective reasons to fall back on. It keeps the bar high and reduces the chance of bias creeping into your decisions.

Your scorecard should:

• Categorize skills into must-have, should-have, and nice-to-have.

• Include both hard skills (technical abilities) and soft skills (communication, problem-solving).

• Limit criteria to a maximum of 12 - more than that makes the process too complex.

Each mission or responsibility in the role should be linked to a corresponding question on the scorecard. Always ask the same structured questions to every candidate. This eliminates random judgment and makes comparisons more reliable.

Here’s a scorecard template we’ve used for Founding Engineer positions:

Case study

Founders often ask us: “How do I evaluate someone in a role I don’t fully understand?” The answer is ideally, you should have done the job yourself first. Whether you’re a CTO who’s built the first version of the product, or a CEO who’s already closed the first sales, having hands-on experience is key. It gives you the context to know what good looks like and what to ask for when hiring.

That’s why the case study you assign should be based on something you’ve already tackled yourself. This allows you to compare the candidate’s thinking and output with your own experience.

Importantly, a case study is never about getting free work or outsourcing an open question you’re curious about. Its purpose is to evaluate how someone approaches real, relevant problems, not to teach you something new.

And if you’re still unsure how to judge the outcome, don’t hesitate to bring in people from your network who have expertise in that specific area.

“A business case is the best way to see how someone thinks and to simulate what it’ll be like to work with them. You give them a real-life problem and see how they approach it. I don’t need someone who has all the answers but I want to see how they think, what questions they ask, where they acknowledge the limits of what they know, and how they start collaborating with us.”

– Evan Testa, CEO and Cofounder of Roundtable

Hiring entrepreneurial mindsets

Evaluating whether someone has an entrepreneurial mindset is one of the hardest parts of early-stage hiring. You need people who think and act like builders but spotting that mindset is tricky, especially when candidates know how to say all the right things.

The key is to ask the right questions and to ask them at different points in the process. You’re not trying to check a box in a first call. You’re looking for consistency over time and stories that hold up under deeper scrutiny.

Here’s how to cut through the noise and figure out who actually has the entrepreneurial DNA your startup needs:

What to ask

• “What other opportunities are you exploring right now?”

  → Are they targeting early-stage companies or defaulting to what’s available?

• “If you had a magic wand, what would your next adventure look like?”

  → If they say they want to be head of engineering at a huge scale-up, it’s a clear sign they’re not interested in a role where they’ll need to build everything from scratch.

• “Why are you looking right now?”
  → If the answer is purely about compensation, titles, or frustration with their current manager, that’s a red flag. You want someone who’s looking because they’re ready to build.

• “If you had two offers tomorrow, how would you decide?”

  → Are they driven by having an impact or chasing prestige? If they say, “I’d go with the bigger name or highest pay,” they’re not thinking like an entrepreneur.

• “Tell me about a time you led an initiative that has had an impact?”

  → Even if a candidate hasn’t started their own company, you can assess their entrepreneurial mindset by asking about initiatives they’ve launched, specifically ones they conceived themselves that had measurable impact on the business or product.

• “Why did you leave your last role? How did you choose the next one?”

  → You can learn a lot by asking this question and uncovering people’s inner motivations.

What to avoid

• Don’t ask vague or leading questions
  Example: “How do you handle risk?” or “Do you see yourself as entrepreneurial?”
  → Of course they’ll say what you want to hear.

• Don’t over-index on startup logos on a CV
  → Just because they worked at a startup doesn’t mean they thrived there. And not all startups are equal. Ask what they did and why they joined.

“The bar we used for early hires was simple: if they left to start their own company one day, would we invest? If the answer wasn’t a clear yes, we didn’t make the hire.”
– Hugues Renou, CEO and Cofounder of Tengo

Do not skip reference checks

Founders often say, “If candidates are giving us the contacts, they’ll only say good things.” That’s a myth. In our experience, reference calls give you a rare, honest window into how someone actually works. You could learn more in 20 minutes with a former manager than in three rounds of interviews, if you ask the right questions.

“During reference checks, you shouldn’t just ask open-ended questions like, “What is this person good at?” or “Not so good at?” Those kinds of questions often lead to vague or rehearsed answers. Instead, I like to use more structured questions, sometimes even multiple choice. I frame it by saying, “Which of these best describes this person: Are they a doer? Creative? Strategic? More of an executor?” We need all types of profiles in the company. But what matters is understanding which profile the candidate fits into — because depending on the role, we may need one more than another.”

– Evan Testa, CEO and Cofounder of Roundtable

Here’s the script we use for every reference call:

Intro & context:

The role is [briefly describe: e.g., a founding engineering position in a fast-paced startup environment]. The company is [stage – e.g., early-stage / scaling], and the role involves [key challenges, e.g., ownership, working across product/tech, autonomy, etc.].

The goal of this conversation is to better understand how [Candidate’s Name] operates and how suited they are to working at a very early-stage company.

1. Contact and context of collaboration

   When did you and [candidate’s name] work together? What was your working relationship (e.g., manager, peer)? What kind of project(s) were you collaborating on?

2. Strengths

   Can you share an example of a project or situation where [candidate’s name] really excelled?
   What stood out about their contributions?

3. Areas for improvement

   On the flip side, was there a situation or project where they struggled a bit?
   How did they handle it, and what did they learn?

4. Collaboration style

   How would you describe their communication style? How do they typically react to feedback?
   Do they actively give feedback to others?

5. Relative performance

   Compared to others in similar roles you’ve worked with,
   how would you rank [candidate’s name]?
   Follow up: so are they in your top 5, top 3? Why? (For example, would you place them in the top 10%, top 25%, etc.?)

6. Fit for a fast-moving environment

   How do you think they’d do in a fast-paced, less-structured

   environment where priorities shift and autonomy is key?

7. Rehire question

   If you had the opportunity, would you work with

   [candidate’s name] again? Why or why not?

8. Conclusion

   Is there anything else we haven’t covered that you think we should know

   about [candidate’s name]?

9. Overall recommendation

   On a scale of 1 to 4 (1 = no, 4 = absolutely), how strongly would you re- commend [candidate’s name] for this role?

Managing the hiring timeline

You need to run a tight hiring process. The longer it drags on, the more likely you are to lose great candidates.

A common pitfall is the case study phase. When founders wait to create the case study until they’re already in the hiring process, it causes unnecessary delays.

There are two hiring metrics you should always track:

1. Time to hire

This is the time from first contact (actually speaking with the person and not the first time you send them a message) to offer acceptance.
→ If this takes more than 2 to 3 weeks, you’re moving too slowly. Every extra day increases the risk of losing top candidates.

2. Time to fill

This is the total time from when a role is opened to when it’s filled.
→ If this goes beyond three months, something in your process is broken. Maybe the bar is unclear. Maybe the pipeline is weak. Either way, it needs fixing.

“Early on, I lost a great candidate because it took us too long to move through the process. After that, we completely changed our approach: we managed to hire someone else in just three days. The idea is simple: you must move at the candidate’s speed. If they’re ready now, you need to be ready now. I remember reviewing a candidate’s test late at night in a bar, laptop in hand, and setting up an offer meeting for the very next day.”

– Paul Vidal, CTO and Cofounder of Collective

Candidate experience matters

You might think candidate experience is a nice-to-have but a sloppy or disrespectful process can hurt your reputation and cost you future hires.

• Always close the loop professionally. If someone reaches final rounds, call them. Don’t send a generic email. It’s a small gesture that signals respect.

• Think long-term. A rejected candidate today might refer a future team member, become a customer, or reapply when they’re a better fit.

• Handle rejections with care. For inbound applications, a clear and respectful rejection email is enough. But for any candidate you’ve interviewed, especially more than once, always call. Walk them through your reasoning, using your scorecard to support the decision. It helps them grow, and it shows you care, which people remember.

How to retain and part ways with employees

Striking the right balance with salaries

At the pre-seed or seed stage, it’s rarely realistic, or even wise, to offer salaries on par with well-funded scale-ups. Instead of overpaying, early-stage startups should focus on offering meaningful equity and a deeply empowering work environment. A real sense of ownership and growth potential often resonates more with the kinds of people you want in the early days.

To ensure fairness and consistency, it’s important to avoid “à la carte” compensation where each person negotiates their own deal. Not everyone negotiates equally, and ad hoc decisions often lead to misalignment and frustration. Even a simple salary grid - based on levels of seniority and scope of responsibilities - brings structure. It also simplifies hiring conversations and scales better as the team grows.

When setting salaries, resist the urge to rely solely on a candidate’s expectations or demands. Just because someone asks for a certain amount doesn’t mean you should pay it. Use your internal framework to guide decisions, that’s what keeps compensation coherent across the team. At the same time, be careful not to underpay someone just because they ask for less. Even if a candidate is willing to accept a lower salary, compensating them well within your defined range avoids future frustrations.

Ultimately, you’re not trying to win people over with high pay. The right early team members join not because of salary, but because of the impact they can have and the long-term upside they see. Well-structured stock options are a great way to share that upside, but they only work if you take the time to explain them properly and tie them to a compelling company vision.

How to think about stock options

Equity is a topic every founder must fully understand and take ownership of. While using benchmarks is helpful, there’s no substitute for informed decision-making. You should talk to peers, align closely with your cofounder, and think carefully about what you want to build in terms of culture, team alignment, and incentives.

Giving equity to your team serves multiple purposes. At early stages, it’s the best way to align economic incentives across the company and attract high-quality talent. Given limited cash, you can’t offer salaries on par with later-stage, well-funded scale-ups. Equity helps close top candidates and provides long-term upside far greater than a competitive salary. It also plays a critical role in retention, aligning long-term commitment and shared value creation through structured vesting over time.

“For us, especially early on, one of the key levers has been stock options. That’s definitely a foundational element and one that retains talent. But just as important is the perspective for growth: giving each person a clear view of how they can evolve within the company.”

– Evan Testa, CEO and Cofounder of Roundtable

Timing is everything

One of the trickiest questions is when to start offering equity to your team. At the pre-seed stage, we generally recommend avoiding any commitment to a fixed number or percentage. Before the seed round, the company is still taking shape, you often don’t know yet how much equity you want to allocate, what the team structure will be, or who will be staying for the long run. Performance differences will soon become visible, and early team members may leave during their trial period. For all these reasons, avoid, if possible, committing specific figures too early.

However, in some cases, especially with high-caliber candidates, you may need to commit to a percentage or equity value to close the hire. That’s fine and there are thoughtful ways to handle this, which we cover further below.

So when is the right time to start issuing equity? In most cases, we recommend doing this after the seed round. The main advantages are clarity and simplicity, you avoid creating an initial ESOP (Employee Stock Ownership Plan) structure before your cap table is stabilized. The downside is that very early employees may pay a slightly higher exercise price. In some cases, where reaching the seed round takes time and a solid “core team” is already in place, it may be worth implementing a small plan earlier.

When hiring early employees, it’s essential to set expectations clearly. Reassure them that equity is a key part of your company culture and that everyone will be aligned through ownership. Let them know that equity grants will follow the seed round, in line with common market practice. Be transparent about the timeline (e.g., equity to be granted post-seed, potentially with retroactive vesting). Share examples and benchmarks from other startups and highlight success stories from your own ecosystem. After the seed round, you can begin to communicate the equity grant amount and vesting start date when making offers to new hires.

Who should receive equity and how much?

We generally recommend allocating equity to all team members within the first 30 to 50 hires. After that, the policy can evolve based on your compensation strategy. Some companies continue to grant equity broadly; others reserve it for key roles, opting instead for higher salaries.

At seed stage, most companies plan to provide somewhere between 5% and 10% of the post-seed cap table to the ESOP. This varies depending on founder count and company valuation. Early employees, typically the first 5 to 10 hires, often receive more grants expressed in % of ownership, without a real link to their salary but strongly linked to their performance. This early group might collectively represent one-third of the ESOP pool. Beyond that, equity starts

in general to be allocated based on benchmarks tied to role type, experience, and impact. A useful framing is linked to the base salary:

60–130% of annual base salary in equity value for senior profiles 40–60% for junior hires

Vesting, grants, and administration

Standard vesting terms are 4 years with a 1-year cliff, which means an employee must work for one full year before any of their equity grant begins to vest. Grants are typically made with vesting starting from the employee’s date of hire. To simplify administration, we recommend grouping vesting events quarterly or semi-annually rather than on a rolling basis. Monthly vesting is generally preferred to avoid “cliff effects.”

Exercise windows should also be well-structured, typically once or twice per year, with a one-month exercise period. This prevents ad hoc exercises and keeps things organized.

Some founders introduce “equity sliders” allowing employees to trade salary for additional equity. While this approach can work, especially when cash is available and salary competition is intense (e.g., for developers), it introduces complexity and scalability issues. Equity is long-term and illiquid; salary evolves over time. Mixing the two can quickly become difficult to manage.

Key rules and planning

Never introduce blanket acceleration clauses, which allow employees to receive unvested equity shares before the normal schedule. Accelerated vesting should be discussed individually and approved by the board on a case-by-case basis for key individuals only.

When planning your ESOP, define your total pool first. Then build a distribution model and align it with your hiring roadmap up to Series A. Forecast the number of hires and average salaries, ensure your pool is sufficient, and add a 25% buffer.

Create a simple, repeatable rule for equity allocation during the early stages - ideally one or two levels. Over time, evolve into a structured matrix based on role, seniority, and market benchmarks.

A golden rule: don’t over-allocate from day one. It’s far better to start conservatively and offer top-ups later to those who excel and grow with the company. Equity, once granted, is extremely difficult to take back without termination. On the other hand, smart top-up policies reinforce loyalty and reward long-term contribution.

Announcing and communicating the ESOP Plan

The purpose of ESOP is for you to talk about it. The way you explain the plan will matter far more than the exact number of shares. Prepare a detailed document explaining how your ESOP works, including examples of potential upside. Share it widely and update it based on feedback and team questions.

Discuss this thoroughly within the founding team, align on your narrative, and be confident when presenting it. For your first team members, announce the framework company-wide, then follow up with individual meetings to present each person’s grant. Be transparent about future top-ups and share stories of employees who grew into leadership roles and significantly increased their ownership.

Make it clear: there is no negotiation, and the policy is based on values, not percentages. Emphasize that equity outcomes depend on multiple factors— fundraising amounts, company valuation, investor terms—and use illustrative case studies to help new hires understand what’s possible.

For post-seed hires, include equity details directly in offer letters. You can also share a public version of your ESOP explainer to appear professional and transparent, while keeping confidential details internal.

Legal considerations

Employees typically have a 30 to 90-day window to exercise vested options after departure. You may offer longer periods (e.g., 6 months) but avoid overly long durations to maintain cap table control. You can consider tiered windows (e.g., 1 year for 2+ years of service, 2 years for 3+ years), and reserve the board’s right to extend timelines in special cases like friendly departures.

Finally, ensure that the company has a call option on the shares of any departing employee to prevent inactive or unfriendly shareholders from remaining on the cap table.

Onboarding

Great onboarding does three things: it aligns, engages, and enables performance.

First, you need alignment. From day one, new hires should be immersed in your company’s mission. They need to understand not just what you do, but why it matters.

Next comes engagement. Early relationships shape long-term retention. Make sure your new teammate feels welcomed. It’s also critical to invest in team bonding early, whether through team lunches or casual time away from office screens. The first few weeks is when newcomers decide whether they feel part of the team or just adjacent to it.

Finally, you want performance. That means setting clear expectations from the first week: what skills they need to master, how they’re expected to contribute, and what success looks like at one, three, and six months in. Don’t wait to share feedback. Give it often, both positive and constructive.

There are a few key milestones that should anchor your onboarding.

• The first week should focus heavily on integration, making sure the new hire connects with the team and understands how things work.

• At the one-month mark, hold a check-in to exchange feedback and recalibrate if needed.

• By month three, you should be ready to make a clear decision about continuing beyond the trial period.

“During the onboarding period, we conduct weekly feedback sessions. As employees settle in, we transition to biweekly sessions, and later to monthly check-ins. These comprehensive 360-degree feedback conversations always begin with the employee sharing their thoughts with us: What’s working? What’s not? What are the risks they see as we scale — what’s fine now but might break if we add 2, 5, or 10 more people?”

— Evan Testa, CEO and Cofounder of Roundtable.

Being consistent with routines

• Once onboarding is complete, strong management routines are what keep performance and growth on track. These routines don’t have to be complex, but they do have to be consistent. Here are some suggestions of routines that work well:

• The weekly one-on-one. These 30–60 minute conversations are where progress is tracked, blockers are surfaced, and challenges are addressed. While these meetings can touch on personal growth, they should primarily stay operational. Deeper development conversations are better suited for monthly or quarterly check-ins.

• Personal development meeting. Every month, or at least every quarter, schedule a more strategic conversation. This is where you zoom out to assess career goals, discuss progress against role expectations, and co-create a development plan. This discussion should be high-level: What’s going well? What needs improvement? What does growth look like? For structure, some founders use a scoring system to track how a person is performing across their core responsibilities (e.g., 0 to 5). Self-assessments and upward feedback can also be part of this rhythm.

• Feedback should flow continuously. Share praise or course corrections in the moment whenever possible. But also carve out dedicated time in your routines to go deeper. Make asking for feedback a habit, both from you to your team, and from your team to you.

• For sales and performance-driven roles, it’s important to tie these routines to goal-setting. Set quarterly targets and use monthly updates to track progress, adjust strategies, and stay accountable.

“From day one, we focused on empowering people. For example, early on, I was the only product manager working with about ten developers. Instead of handing them detailed specs, I’d sit with them, explain the problem, and ask them to come back with a solution. We’d challenge their ideas until we aligned, but the solutions always came from them. They truly owned the work. The only thing I fully designed myself was the MVP’s API interface, everything else, they built and shaped themselves.”

– Nicolas Saison, CPO and Cofounder of Swan

Spotting red flags

You’ve probably heard the saying “if there’s a doubt, then there’s no doubt”. It’s often that founders make the mistake of keeping someone who’s underperforming for too long expecting different results; You often have to do that mistake once to realize it and learn your lessons, and that’s ok. But here’s a list of some red flags to spot to see if someone is underperforming.

• Low ownership mindset: If they talk only about execution and not about outcomes or impact, that’s a problem.

• “Tell me what to do” attitude: You’re looking for builders, not task-doers.

• Avoids ambiguity: If someone needs everything structured, they won’t survive startup chaos.

For this episode of PMF Collector, Matthieu Vaxelaire sits down with Fabrice, co-founder of Yoobic — a platform used by retail networks to manage store operations, from communication and training to daily execution across hundreds of locations.

Built with his two brothers, Yoobic didn’t start with a product idea.

It started with a conviction: most people don’t work behind a desk — yet the tools built for them are still designed for headquarters.

What follows is not a story about luck or timing. It’s a story about method.

You can watch the full conversation above, or read the key themes below.

1) Starting from a conviction — then forcing it through reality

Yoobic didn’t begin with a feature.

It began with a simple observation: the vast majority of workers are “deskless”, yet almost no software is designed for them. That gap was the starting point.

But instead of jumping into building, Fabrice and his co-founders approached it methodically. First, identify a technological shift — in this case, mobile. Then apply it to a vertical where the gap is obvious.

Retail quickly emerged.

Still, a top-down conviction wasn’t enough. It had to survive contact with reality.

2) The only way to understand a problem: live it

So Fabrice went on the ground.

He spent weeks shadowing store managers, observing how they worked day to day. What he found wasn’t inefficiency — it was chaos. Hundreds of emails from HQ, constant onboarding due to high turnover, fragmented tools, and processes that simply didn’t scale.

In one case, sales reports were still being sent by fax.

“When you’re inside the system, you don’t see how broken it is. From the outside, it’s obvious.”

This wasn’t a feature gap. It was a structural problem.

3) Narrowing down to one painful, recurring process

Instead of trying to fix everything, they made a key decision: go extremely narrow.

They focused on a single process — executing promotions in stores.

Not because it was glamorous, but because it met three critical conditions. It happened frequently, it had a direct impact on revenue, and someone inside the organization was accountable for it.

That was enough.

“If you try to solve everything, you don’t have an MVP anymore.”

So they built only that.

4) The only validation that matters: a paying customer

From the beginning, Fabrice had a clear rule: no free validation.

The MVP had to be paid.

Even if the amount was small, the act of paying forced a different level of commitment. It meant budget approval, internal alignment, and a real problem being solved.

After a short pilot, the signal was immediate.

The client came back and said: “I want this in all my stores.”

That was the first real proof.

“The amount doesn’t matter. The fact that they pay does.”

5) Growth came from focus, not expansion

Retail is a huge market. Too huge.

So instead of expanding, they kept narrowing.

Through repeated sales conversations, they refined their positioning, their messaging, and their target. Eventually, they landed on fashion retail — a segment where operational complexity is high and execution directly impacts revenue.

This focus made everything easier: clearer messaging, faster sales, stronger traction.

“The more precise your target, the easier it is to sell.”

6) Customers as the only real growth engine

In the early days, Yoobic didn’t rely on paid acquisition.

They relied on their customers.

Case studies, testimonials, conferences — in a small ecosystem like retail, word spreads quickly. Once a few brands adopt a solution, others follow.

This created a compounding effect.

“Your best marketing is your customers.”

7) Enterprise sales: slow, but predictable

Selling to large retail networks meant dealing with long cycles — often six to nine months.

To make this work, Yoobic built a structured approach around proof. Start with a limited deployment, measure the impact, and turn that into a quantified business case.

Once the ROI is clear, the conversation changes.

“If you can prove ROI, you don’t need to wait for next year’s budget.”

8) The US expansion: rebuilding everything except the product

After finding traction in Europe, Yoobic expanded to the US.

But almost nothing transferred.

No brand, no references, no network. So Fabrice made a deliberate decision: start from zero again. Small team, narrow ICP, outbound-driven approach.

The product didn’t change.

Everything else did — from messaging to pricing to the sales process.

“Just because it works in Europe doesn’t mean it works in the US.”

It took months to land the first clients. Then momentum came back.

9) A company is built through iterations, not plans

At every stage, the approach stayed the same.

Form a hypothesis, test it quickly, define success in advance, and iterate. Short cycles. No endless discussions.

First, open debate. Then, once a decision is made, full execution.

“You need both: open debate first, then ruthless execution.”

This discipline compounds over time.

10) Keeping the founder mindset at scale

Today, Yoobic is a large organization.

Which creates a tension: you can’t operate like an early-stage startup anymore — but you can’t lose that energy either.

Fabrice’s solution is to carve out small spaces to keep experimenting. Not on the core business, but on the side. Fast loops, quick iterations, then reintegrate what works.

“If I do this on the core business, I break the company.”

It’s how he keeps the “day one” mindset alive.

11) Persistence vs. stubbornness

Behind everything, there’s one constant tension.

You need persistence to push through uncertainty. But too much of it becomes stubbornness — and keeps you going in the wrong direction.

The real skill is knowing when to adjust.

“You need to keep going — but not in the wrong direction.”

This episode is a masterclass in building vertical SaaS from the ground up: starting with a real-world problem, validating through payment, focusing relentlessly, and rebuilding your playbook at every stage.

This chapter is about scoping, building, and iterating your Minimum Viable Product (MVP), the first version of your product that lets you test your core hypothesis in the real world. The goal isn’t to build something polished or complete. The goal is to launch fast, get it in front of users, and learn as much as possible.

Many founders fall into a common trap: building too much, too soon. It’s tempting to perfect every feature, but most of them won’t matter or won’t even be used. Worse, you risk missing your moment: by the time you launch, the market might have shifted or your solution may no longer fit the problem.

At Hexa, we believe you should ship your MVP within weeks, not months. In this chapter, we’ll show you how to scope tightly, build quickly, and iterate with design partners.

A note: this chapter was originally written in 2025. Since then, tools like Cursor and Lovable have compressed the path from idea to working product from weeks to days. These tools are incredibly useful, but it also makes it easier than ever to overbuild. Arguably, the principles in this chapter matter more now than ever.

Scope small but think long-term

Every great product starts with a vision, a clear idea of the impact it will have. But too often, founders get lost in execution and forget to ask: what are we ultimately building?

Your MVP isn’t the final product, it’s the first step. It’s not about cramming in features but proving your vision.

Start with your long-term vision

Before diving into building your MVP, take a step back and define your long- term vision. What are you ultimately trying to achieve with this product? Your MVP is just the first step in making that vision a reality, the way to test your core hypothesis, not the finished product.

Think long-term. You’re not just building for today; you’re laying the foundation for something bigger.

Once your vision is clear, ask yourself: What’s the first step toward achieving it? This first step should deliver enough value for users to engage with your product and use it regularly. It should solve one fundamental problem. Rather than trying to do everything at once, focus on perfecting your core use case, the one that provides the most value to users right now.

Your MVP = the first step toward your vision

Your MVP may only represent 1% of your ultimate vision, but it should deliver 80% of the core value.

Examples of long-term visions and MVPs

Below are the initial long-term visions of these companies. It’s important to remember that long-term visions often evolve, they’re not set in stone. Just like Aircall, Spendesk, and Front, most startups adjust their vision as the market shifts. While your vision shouldn’t change every week or quarter, it’s still a hypothesis - one that needs to be tested and refined over time.

Aircall

• Long-term vision: A modern cloud-based phone system that integrates seamlessly with business tools (CRM, helpdesk, etc.).

• MVP: A simple VoIP solution that allows businesses to set up a cloud-based phone number in minutes. The first version had basic call routing and a web- based dialer, no deep CRM integrations, analytics, or call center automation.

Spendesk

• Long-term vision: An all-in-one spend management platform that replaces corporate credit cards, expense reports, and manual approval processes.

• MVP: A virtual card solution for employees to request and manage company spending. The first version had basic expense tracking and approval work- flows, no complex reporting, invoice management, or budgeting features.

Front

• Long-term vision: Reinvent email for teams, making shared inboxes more collaborative and efficient.

• MVP: A shared inbox for customer support teams that allowed multiple people to manage emails together. The first version had basic email assignment, internal comments, and tagging, no deep integrations, analytics, or automation.

Launch your MVP in weeks, not months

You might wonder: Why scope small? Why not deliver an MVP that includes the first two steps of your long-term vision? Why be so ruthless about delivering the smallest MVP with the biggest impact?

There’s one reason, and one reason only: you need to ship fast. The sooner you get your product into the hands of users, the sooner you’ll know if your first step truly delivers value. Scoping small allows you to launch quickly, gather real-world feedback, and validate whether you’re on the right track. No internal discussion or assumption can replace the insights gained from real user interactions.

The goal is simple: get it in front of users in a matter of weeks, not months. You’ll learn 10x more from real usage than from endless internal debates.

Plus, speed is everything in startups. One of the most common mistakes founders make is spending six months or more perfecting an MVP before launching. That’s the wrong approach. In a startup, speed is your greatest advantage. Speed means getting real feedback before you’ve invested too much time in the wrong direction. It means staying ahead of competitors, maintaining urgency within your team, and proving your ability to execute.

But there’s an important caveat: not all B2B software MVPs can be launched in a matter of weeks. While many can, and arguably should, some industries make that timeline unrealistic. Highly regulated sectors like fintech often require long lead times. For example, Swan had to wait 18 months to obtain its e-money license before launching. In other cases, the nature of the product demands a longer build cycle. If you’re entering a mature category with high user expectations (like email, calendar apps, CRMs, or ERPs) you may need to build for months before reaching the level of polish and value users expect. These are categories where the bar is already high. That was the case for us at Folk, and similarly for Front.

A common failing of founders

One of the biggest mistakes founders make is skipping the scrappy MVP and instead spending time crafting a polished prototype that reflects their long-term vision. They show this prototype to potential users, hoping for validation.

Users do not react to a prototype the same way they do to a real, functional product. A prototype invites opinions, but a working product triggers real behavior. If you rely on feedback from a non-functional prototype, you risk making decisions based on theoretical reactions rather than actual usage.

It’s best to skip the theoretical stage. Instead, build a minimal but functional MVP, something people can truly use. Only then will you gather real insights, understand how users interact with your product, and iterate in the right direction.

The hammer allegory

As a founder, it’s tempting to wait until your product matches your full vision before sharing it with users. You imagine launching something beautiful and complete - a product that truly reflects your ambition. But you can’t deliver that perfect version yet. You don’t have enough feedback or iteration behind it. And ironically, the only way to get there is by putting something far less perfect in users’ hands today.

Let’s take the example of a hammer as a product.

Imagine someone trying to drive a nail into a piece of wood using just a rock. It’s clumsy and frustrating. You immediately picture the solution: a hammer. That’s your long-term vision.

But here’s the reality: you can’t hand them the final hammer today. You don’t yet know what the handle should be made of, what weight the head should be, or how it feels in actual use. So instead, you have to start smaller.

Day 1: You hand over a cardboard prototype. It resembles a hammer, but it’s useless in action. It bends. It fails. It leaves you with no insights and them with no progress.

Day 2: You build something crude but real: a stick with a rock tied to one end. That’s your first step to reach your long-term vision. Yes it’s ugly and unpolished. But it’s better than the rock alone. They use it and it works, barely, but now they’re engaged. “Can this be longer?” “It needs to be heavier.” Suddenly, they’re helping you build the hammer.

That’s the point: early, imperfect tools generate real feedback. Real usage reveals real needs. Every rough edge is a learning opportunity you wouldn’t get from polish alone.

You can’t ship the hammer today, but you can ship something usable. Not to validate the final product, but to start discovering what it actually needs to become.

It’s not about lowering the bar, it’s about laying the foundation. The only way to build the hammer is to first hand someone a better rock.

Your wireframes should be ugly

When founders start working on their MVP, there’s a natural tendency to overcomplicate, to add more features, more customization, more flexibility. But in reality, complexity dilutes value. The best products in the world share one common trait, they are deceptively simple. Slack, Notion, Figma, all of them took complex ideas and distilled them into an intuitive, frictionless experience.

But here’s the catch: to build simple products, you have to start simple. And that begins with low-fidelity wireframes.

Starting with low-fidelity wireframes is crucial because it allows you to:

Iterate fast

Wireframes are quick and cheap to modify. If an idea doesn’t work, you can sketch a new version in minutes rather than reworking high-fidelity designs or code.

Simplify your MVP

Most MVPs start with too much. You should be aggressive about removing features. Cut everything in half: half the buttons, half the options, half the tabs. Low-fidelity wireframing keeps things scrappy and ensures you can cut it all out.

Save time & resources

Designing in high fidelity (detailed UI, colors, animations) takes significantly more time. If you need to make changes later (which you will), it’s much easier to adjust a wireframe than to rework a polished design or rewrite code. This reduces wasted effort.

Collaborate with your team

Low-fidelity wireframes are easy for everyone to contribute to, whether they’re designers, developers, or product managers. Since they’re not overly polished, team members feel more comfortable suggesting changes and iterating together.

Reduce your bias toward aesthetics

When something looks finished, people hesitate to challenge it. With low-fidelity wireframes, stakeholders focus on functionality and flow, rather than prematurely debating colors, typography, or UI details.

Top tip: Use Blocks.pm for low-fidelity wireframes.

Hexa created a Figma plug-in to create low-fidelity wireframes as a team. It has tens of blocks ready to go, for you to move around and edit as you need.

Spendesk’s early wireframe - made on Blocks

Your MVP building checklist

• Aligns with your long-term vision – It should be a stepping stone toward your ultimate goal, not a random experiment.

• Can be developed in weeks, not months – If it takes too long, you’re over- building.

• Is a real product and not a polished prototype - The only way to get mea- ningful insights

• Delivers enough value to onboard real users – Even in its simplest form, it must solve a real problem.

• Is minimal enough to allow for quick iteration – No unnecessary complexity that slows future changes.

If your MVP plan doesn’t check all these, you’re overbuilding.

Think in milestones, not roadmaps

Make your planning flexible

When you’re building a startup, long, detailed roadmaps can be more of a burden than a guide. Plans change and rigid timelines can slow you down instead of helping you move forward.

At Hexa, we believe in momentum over meticulous planning. Instead of a complex roadmap, focus on short, achievable goals that drive real progress and bring you closer to your long-term vision.

Forget the Gantt charts, the multi-quarter planning, and the overly detailed feature breakdowns.

Your roadmap should be a simple table. Here’s a suggested template:

Here’s an example of a roadmap we have for an internal project called LegalX:

Product rituals

Even if you’re just one or two founders, establishing a clear rhythm for product discussions, prioritization, and execution will keep you moving efficiently. The key is consistency. These rituals help you maintain focus, iterate faster, and ensure that small, incremental progress compounds over time.

Here’s what early stage product routines should look like:

1. Weekly product committees

Cadence: Every week, same time.
Duration: 30 minutes to one hour (keep it fast and focused).
Who’s involved: Everyone working on the product (even if that’s just you).

What happens?

• Review what was shipped: every feature, bug fix, or improvement from the past week should be demoed (even if it’s rough).

• Refine priorities for the upcoming week: what’s the one most important thing to focus on next?

• Address blockers: anything slowing you down?

Why it works:

• Establishes a consistent cadence for execution.

• Prevents features from getting stuck in endless iterations.

• Forces you to ship, because you have to demo something every week. Keeps the team aligned to achieve quarterly milestones.

2. Quarterly milestone planning

Cadence: Every 3 months.
Duration: 1-2 hours.
Who’s involved: Founders, product team, and key stakeholders.

What happens?

• Define big product milestones for the next quarter.

• Align on the core problem to solve and major initiatives.

• Set clear execution goals that break down into weekly sprints.

• Discuss what’s working and what needs to change in the product approach.

Why it works:

Prevents you from getting lost in day-to-day execution. Keeps you on track to achieving your long-term vision. Helps you track progress and adjust course if needed.

It’s easy to think that formal processes like these aren’t necessary in the early days. “We’re small, we talk all the time anyway.”

But without a structured cadence, things can quickly become chaotic:

• You lose track of priorities.

• You keep iterating endlessly instead of shipping.

• You don’t reflect on what’s working and what’s not.

By locking in these rituals early, you create a strong execution rhythm that scales as your team grows. They also make it easier to stay agile and pivot if necessary, because you’re constantly reflecting and refining.

Your goal:

Think in 3-month cycles: plan quarterly milestones. Execute in weekly sprints: ship and adjust every week.

How to validate early PMF

Why design partners matter

Launching your MVP is just the beginning, the real challenge is proving your product truly adds value to users.

At Hexa, we initially relied on pilots (beta testers) for early-stage product testing. However, we’ve since switched to working with Design Partners (DPs), an approach that transforms early users into active co-builders of our products. In Chapter VI, we’ll go much deeper into how to run Design Partner programs. For now, here’s a quick introduction to why they’re so valuable at the MVP stage.

Our old pilot programs gave early, free access to users, often with little commitment. If they liked it, great. If not, we moved on. However, this passive approach rarely provided clear signals that we were moving toward product-market fit.

A Design Partner Program changes that. Instead of simply testing usability, design partners actively co-build the product with you, ensuring it solves a real problem within a real workflow. This approach requires more effort and commitment from both sides, but the value it delivers is enormous.

Beyond gathering feedback, a Design Partner Program serves as an early validation of PMF. We emphasize the “early validation” part because achieving true PMF won’t happen in the first 12 months. On average, B2B products take over two years to reach PMF.

But a design partner program allows you to test whether users truly need your product, not just whether they find it interesting. If you struggle to find committed design partners, it’s an indication that your product might not yet address a strong enough pain point. Conversely, if companies are eager to participate and actively shape the product, it’s a signal that you are onto something valuable.

Pilots vs. DP program

Co-creating with DPs

We’ll deep dive into how to find design partners in the dedicated chapter on how to acquire your first users later in this book. For now, here are some pointers on how to best engage them.

A Design Partner Program isn’t just about collecting feedback, it’s about embedding users into your product development process. Here’s how to do it effectively:

Daily conversations with users

Regular check-ins help you gauge if your product is truly solving their problem. Design partners should feel like they are part of the team, and their input should influence your roadmap.

Iterate with every prototype

Instead of waiting for big launches, share each iteration early and often. Every version should be tested by design partners to validate the changes before broader deployment.

Observe real behavior

Actions speak louder than words. Don’t just rely on verbal feedback, watch how users interact with the product to uncover friction points and unexpected use cases. Watching users interact with your product often reveals insights they can’t articulate in interviews.

Create a dedicated space for collaboration

Use a Slack channel or private forum where design partners can:

• Share thoughts and feedback in real-time

• Stay updated on product changes

• Feel engaged as co-creators

Asking the right questions

To make the most of your design partners, ask open-ended and practical questions:

• “What problem were you trying to solve when you used this product?”

• “How are you currently solving this problem without our product?”

• “What would make this 10x better for you?”

Good questions should:

• Be open-ended - Avoid yes/no answers; let users elaborate.

• Be short & direct - Simple questions like “Why?” or “How?” drive deeper insights.

• Prompt demonstrations - Ask users to show how they interact with the product, as behavior often differs from what they say.

Analyzing user behavior

Once feedback starts coming in, it’s essential to dig into usage patterns:

• What features are used most often?

• Where do users drop off or struggle?

• Are there surprising use cases we didn’t anticipate?

This analysis will highlight what’s working, what’s not, and where to focus improvements.

Managing feedback without losing focus

With so much feedback flowing in, there’s a temptation to add every feature request. But blindly saying yes leads to a bloated and unfocused product. Instead, filter feedback carefully:

• If users say “this isn’t good enough yet”, it’s a sign they see potential.

• Prioritize feedback that aligns with your long-term vision

• Store all feedback in one place, every suggestion should be considered, even if it’s implemented later.

I launched my MVP, now what?

Focus on one metric

When you get started it’s tempting to measure every corner of your product. The most successful teams use a “focus” metric, a metric that can only move when your business grows in a healthy way. Once you have a focus metric you can use it as a compass to justify your decisions and to try and understand if you’re on the way to product-market-fit or not. This doesn’t mean you should focus on just one metric (tracking several is valuable) but it’s important to have a single north star to guide your decisions.

For B2B software, a focus metric can be one of the below:

Customer retention rate

The % of users returning after a set period. Retention is everything for early-stage startups. A strong retention rate signals that your product delivers ongoing value. If retention is low, if your design partners are dropping off, then it’s a red flag.

Monthly recurring revenue (MRR) growth

We’ll cover the specifics of this in the acquiring clients chapter. MRR is another key metric to focus on in the early stages. In fact, monetizing your MVP early can be one of the strongest signals of product-market fit. When design partners are willing to pay even for a minimal product, it’s a very positive indicator for the future.

Active users

Whether that’s daily/weekly activation rates, tracking active users is crucial because it directly reflects engagement and product stickiness, helping you gauge whether users find real value in your product. Unlike vanity metrics like total sign-ups, active users show how many people consistently return and interact with your product. A growing and retained active user base signals that you’re moving toward PMF.

But don’t spend too much time on it either

Yes, tracking one of the above metrics is important but do not spend excessive time tracking everything. Early-stage product development should be driven by conviction and qualitative feedback, not excessive data analysis. Talking to users is far more valuable than getting lost in dashboards. Instead of obsessing over metrics, focus on:

• Building a product you would like to use*

• Observing user behavior through session recordings

• Talking to users, everyday

• Using the above analytics to validate, not dictate, decisions

*while this may seem to conflict with solving users’ problems, it’s crucial because you’re ultimately the judge of quality. If you don’t believe in your product, how can you expect others to?

Should I pivot?

It’s a big question and one where the framework we introduced in Chapter
II becomes especially useful. As we discussed, a strong startup idea is built on three core components: a problem, a solution, and a GTM strategy. When considering whether to pivot, revisit each of these elements. If any one of them falters (if the problem isn’t painful enough, the solution doesn’t resonate, or the GTM strategy isn’t delivering) then a pivot may be the right move.

One important thing to keep in mind, and why the startup idea framework is so useful, is that you might be solving a meaningful problem with a solution users truly love, yet still struggle because your GTM strategy doesn’t scale or your business model doesn’t capture enough value. That was the case with Collective, who pivoted post-seed, from a hybrid SaaS/service model to a social network-style platform for freelancers.

Insight from Simo Lemhandez, CEO & Cofounder of Folk

Early on, we used a mix of qualitative and quantitative metrics:

1. Qualitative metrics:

We ran surveys like “How disappointed would you be if you could no longer use the product?” (from Superhuman), and standard tools like NPS. These gave us early insights into user satisfaction and perceived value.

2. Engagement metrics:

We tracked usage ratios like:

• WAU/MAU (Weekly Active Users over Monthly Active Users)

• DAU/MAU (Daily Active Users over Monthly Active Users)

The underlying assumption here is: the more intensely and frequently people use a product, the more valuable they perceive it to be.

Of course, this depends on the product category. For example, with a tool like PayFit (HR/payroll), the goal is to use it as little as possible. But for productivity tools like Folk, high usage is a good proxy for value.

3. Talking to users:

This is often underrated, but it was critical for us. We had frequent conversa- tions with users to gather qualitative feedback. Are they satisfied? Frustrated? What pain points are they raising?

One thing I found particularly insightful was the idea that you should aim to create “love” for your product and not neutrality.

Surprisingly, the best way to get there was often to start with users who were frustrated or even a bit angry. Because if they care enough to complain, they see the product’s potential.

The real danger is users who are indifferent, who just don’t care, which means you’re not solving an important problem for them.

In fact, we found it easier to turn “haters” into champions than to convert lukewarm users (like a 5/10 or 6/10) into true fans (9/10). The hate-love spec- trum is useful, because hate shows there’s something worth fixing. Indifference, not hate, is the true enemy of product-market fit.

Insight from Paul Vidal, CTO and Cofounder of Collective

Pivot stories are extremely valuable and we don’t have enough of them, especially in Europe. Too often, we only hear about startups that had a straight, smooth trajectory. But the reality is different: many companies pivot, struggle, adapt, and eventually succeed.

Our initial business was actually doing quite well. By the end of 2024, when we started considering the pivot, we were on track to generate €1M revenue when we started thinking about the pivot. Many companies would have been happy with that.

But we realized two things:

1. We were building a very operational, service-heavy business. Scaling from €1 million to €10 million would have been slow, costly, and painful.

2. Beyond €10 million, the growth path became even less clear, it wasn’t an exponential model.

We had raised €8 million, but our revenues weren’t scaling as fast as we expected. Our business model was a hybrid of services and tech, which made it even harder. Service businesses require aggressive cost management, while tech companies require heavy R&D investment. Doing both at once was stretching us thin.

And so we pivoted toward a bigger opportunity.

We knew the freelance space extremely well after three years. We had deep insights, strong relationships, and brand recognition among freelancers and companies. We realized there was a gap in the market for better software tools, ones we wished we had when we started.

So instead of continuing to grow a service business slowly, we decided to build the software product we ourselves would have loved to use.

Hexa Product Lead Pierre sits down with Louis Pinsard, CTO & co-founder of Dialog — an AI shopping assistant handling hundreds of thousands of conversations per month across 200+ ecommerce stores — to explore how they scaled from basic RAG to the infrastructure that supports several hundred thousand conversations a month.

They started with a very basic RAG setup, which was enough to land their first client. But pretty quickly they hit the real problems: product catalogs that explode in size and break your retrieval, a growing volume of users with increasingly diverse queries, keeping latency low through all of it, and building evaluations that give you objective criteria to actually know whether your agent is getting better or worse.

Louis walks through the full technical evolution, from chunking strategies and retrieval tricks to the evaluation systems they wish they’d built sooner.

You can watch the full conversation above, or read the recap below.

Subscribe now

1) The first version was literally a LangChain tutorial in production

Dialog started in early 2024, right around when GPT-4 had been out for a couple of months. The team needed to move fast, so they did the most basic thing possible: a textbook RAG setup. User sends a message, you search for semantically close documents, stuff them into the LLM’s context, generate a response. A few lines of code, shipped to production.

It worked well enough to get about ten Shopify design partners on board. There was no real evaluation system — Louis would test a handful of queries manually and go with his gut. The models were also expensive, but the bet was that prices would keep dropping. That bet has held up.

2) Not every message deserves an expensive LLM call

Pretty quickly, they noticed people typing “hello” or sending off-topic messages, and the system would dutifully burn through tokens on zero-value interactions. So the first meaningful addition was a routing classifier that decides what kind of message this is before anything else happens. Is it store-related? About a specific product? A policy question? Or just noise?

This became one of the most important components in the whole system. The core idea hasn’t changed since: not all queries should be treated the same way, and sorting that out first saves money and improves quality simultaneously.

3) Re-ranking fixed the hallucination problem

Hallucinations came from two places: the retrieval step failing to surface the right documents, or surfacing too many and the LLM getting lost in an oversized context.

The fix was a re-ranking step. Instead of retrieving 10 documents and hoping for the best, you first cast a wide net — about 100 candidates — then run them through a specialized model whose only job is to pick the 10 most relevant ones for the given question. Tighter context, better results, fewer hallucinations.

4) The user’s question is almost never good enough to search with directly

If someone is on a skincare product page and asks “is this good for my skin type?” — that query has almost no useful semantic content. No product name, no ingredient, nothing for retrieval to latch onto.

One technique that worked well early on is HyDE — Hypothetical Document Embedding. You ask the LLM to generate a plausible answer without context, then use that hypothetical answer as your search query. Even a hallucinated answer will be semantically closer to the real documents than the vague original question was.

They also enrich queries with page context, conversation history, and keyword extraction. The rephrasing step turned out to be one of the highest-leverage improvements in the whole pipeline.

5) How you chunk your product catalog matters more than you’d think

A product has a title, a description (sometimes enormous), variants, collections, metadata — and if you naively chop all of that into fixed-size chunks, you get fragments where the end of a description bleeds into collection data. That’s meaningless to a retrieval system.

Dialog splits along logical boundaries first — description separate from title separate from variants — then applies size-based chunking within those sections. They also attach metadata like the product name to every chunk so fragments can always be traced back to their source. This becomes critical when you need to filter by price, skin type, or other attributes.

6) The filter trick that made structured catalogs useful

Many ecommerce catalogs already have structured filter systems — skin type, brand, price range. Dialog realized they could extract filters from the user’s natural language query before even hitting semantic search.

So “what’s a good moisturizer for dry skin?” gets “dry skin” extracted as a filter, the search is restricted to matching products, and semantic retrieval runs within that smaller subset. It’s adding determinism into a fuzzy process — and it works because the catalog structure already exists, you just need to tap into it.

7) They flew blind on evaluation for the first six months

For the first six months, improvements were measured by vibes. Louis would test manually, look at logs, ask clients to try things. That worked while the wins were obvious. But eventually changes got subtler and the risk of breaking something elsewhere got real.

The classifier was easy to evaluate — known input, known correct category, score it. RAG evaluation is much harder. You need query-document pairs that represent ground truth, and building that dataset requires genuine domain expertise. They use a mix of LLM-as-judge and human review, but Louis is clear this is still evolving.

8) What Louis would tell someone building an agent from scratch today

Three things.

Observability early — log the inputs and outputs of every component from day one. You’ll always be glad you have the data later, even if you don’t know exactly how you’ll use it yet.

Don’t jump to vector databases. Classical text search — BM25, Elasticsearch, even grep-style approaches — works better than people think, especially with structured data like product catalogs. Louis points to Claude Code as an example: it searches codebases through clever grep queries, not embeddings. Start simple, add semantic search when you’re genuinely hitting limits.

And use Python if the AI system is the core of your product. The ecosystem, tooling, and talent pool are all there. It sounds obvious, but it’s not always the default choice.

If you’re building AI agents on top of messy real-world data, the full conversation is worth your time. Louis gets into the details in a way that’s rare for a CTO of a company at this stage.

Subscribe to receive future episodes of Build with AI.

Two of the most common causes of startup failure are not finding product/market fit and co-founder issues. And while the first gets most of the attention, the second can kill a company before it even gets off the ground.

There’s no universal formula for successful co-founder matching. The origin stories vary, some teams are forged out of long-time friendships, others come together fast around an idea or within a startup program. Some partnerships feel like family, others remain strictly professional.

After matching over 100 founders at Hexa, we’ve seen clear patterns in what works and what doesn’t.

The best co-founder relationships tend to be built on three key ingredients:

1. Complementary skills and personalities
2. Alignment on 5 non-negotiables values
3. Transparency, trust, and communication

This chapter breaks down each of these key ingredients, how to evaluate them early, and how to reinforce them over time so the partnership stays strong.

What makes a great cofounder team

Yes, we think having a cofounder is essential - especially if you’re building a software company. You need both technical and business expertise not just on the team, but at the founding level. In software, the product is the technology. Early decisions about architecture, scalability, speed of development, and user experience can define your company’s trajectory. These choices require deep ownership, something you only get when the person leading the technical side is a true cofounder. Conversely, you need business expertise at the founding level because building a great product isn’t enough, you have to position it, price it, distribute it, and fund it.

Why you shouldn’t be solo

Having a co-founder increases your odds of success because of three reasons:

Reason 1: You can’t be great at everything. No one is. A good cofounder fills your gaps. Between technical, product, go-to-market, people, finance and more, there are a lot of areas of expertise you need to cover. We’ll go deeper on that later, but for now: skill range matters, and you rarely get it all in one person.

Reason 2: Startups are hard and lonely. Building something from scratch is emotionally intense. You’ll have days where you feel unstoppable, and others where you’ll question the whole thing. A cofounder won’t make the chaos go away, but they’ll be in it with you. And that makes a huge difference.

Reason 3: You make better decisions, faster. Every founder goes down the wrong path at some point. You get excited by a new direction, overcommit to an idea, ignore a red flag. The difference is whether you have someone to help catch it early. A cofounder helps you make decisions faster, but also course-correct faster. And that can be the difference between a three-week distraction and a six-month detour.

Two is the sweet spot

For most B2B software companies, having two cofounders is an optimal setup. It gives you the benefit of complementary strengths without the complexity of a larger founding team. One founder brings technical expertise, while the other focuses more on business, operations, or go-to-market, but the most successful pairs co-own the strategy, vision, and major decisions.

Sometimes though, two isn’t enough.

If the business model is more complex, or the first two cofounders don’t fully cover the core skills needed to build and run the company, a third cofounder can help close the gap. This is especially true in operationally intense businesses (like ecommerce or logistics), or regulated industries (like fintech), where operations or compliance are critical business functions.

Take Swan. It was built around three cofounders: CEO, CTO, and CPO. Nicolas Saison joined as cofounder and CPO to lead both Product and Compliance, the latter being essential in a regulated space.

But it only works when each founder has clear and complementary ownership. When that’s not the case, adding a third person can actually create more confusion than clarity.

And while three can be the right number, going beyond this rarely is. Larger founding teams tend to slow down decision-making and blur accountability.

Three ingredients

We believe that a good cofounder relationship has three essential ingredients:

1. Complementary skills and personalities

Founders should bring different strengths to the table. One might be a visionary builder, the other a sharp operator. One thrives on product intuition, the other’s obsessed with numbers. It’s also about personality. Is one more optimistic while the other brings healthy skepticism? The goal is to balance each other out and not be the exact same person.

2. Alignment on the 5 non-negotiable values

3. Transparency, trust and communication

Most co-founder breakups don’t happen suddenly. They develop gradually through misaligned expectations, poor communication, and issues that go unaddressed too long. The best co-founders prevent this by talking regularly, engaging in honest disagreements, and giving direct feedback.

How and where to find one

Expect that it will take time and volume

To find your co-founder, you should think of it as a sales process. You need to actively build a funnel. Most people you talk to won’t be the right fit, and that’s okay.

At Hexa, most founders who successfully find their partner:

• Talk to at least 20 people properly before things click

• Take 2–3 months to find the right person

• Spend 2+ hours of their day on it - treating the search like an integral part of their daily job, not an afterthought

You don’t need a fancy CRM, but if you’re juggling lots of conversations, a simple tracker helps you stay clear on where momentum is building.

On a simple Notion, Google doc or Airtable, track four things:

• Who you’ve contacted

• Who you’ve spoken to

• Who you want to keep exploring with

• Who you want to nurture for your future first hires: because even if someone you meet doesn’t fit your criteria to become a founder, they might make an excellent early employee so do not forget them!

Make your project discoverable

If you’re serious about finding a cofounder, you have to show them what they’d be joining.

Don’t hesitate to post on social media saying you’re looking for a cofounder, especially if you pair it with what you’re building and where you want to go. And no, sharing your idea won’t get it stolen. At Hexa, we’ve always been open about our early-stage ideas when looking for cofounders, and not once has someone run off with one. The fear of being copied is usually much bigger than the actual risk.

Leverage your network

The ideal situation to find yourself in is to team up with someone you already know. Someone you’ve worked with before, collaborated with on a side project, and hopefully even navigated messy problems with. That kind of shared history takes a lot of the unknowns off the table. You already have a sense of how they operate, not just when things are going well, but when things are uncertain or tense.

A lot of these people will already have jobs, but don’t assume they’re unavailable, uninterested, or too far along in their careers. Make the ask. People’s situations shift quickly, and a well-timed conversation can unlock something unexpected.

It’s also worth asking people you admire: “If you were starting something tomorrow, who would you want to cofound with?”

You’ll get names you wouldn’t have thought of, people who’ve earned deep trust.

Spotting entrepreneurial intent

You’ve exhausted your network and still nothing? You can also start by reaching out to people directly. But intent is hard to spot. Most people don’t have “open to co-founding a startup” in their LinkedIn bio. That doesn’t mean they’re not interested, it just means you need to read between the lines.

The key is to look for signals. You’re hunting for mindset as much as skillset. People who’ve built side projects, written about startups, or worked at early-stage companies might have that entrepreneurial itch, even if they’re not shouting it.

You can use targeted keywords to surface the right people and filter for potential cofounder energy.

Here’s a non-exhaustive list of keywords that can be useful in your search to detect that entrepreneurial intent:

Founder: sounds very obvious, but it’s actually even better if they pop up as ‘founder’ in a previous position rather than current. This signals they’ve done it before, know how hard it is, and might be ready to try again. At Hexa, being an ex-founder of a failed startup is a big green flag. It means they’re familiar with the trials and tribulations of entrepreneurship and know what they’re getting into.

Stealth startup: or titles such as “Working on something new.” It usually means they’re in the very early stages. While that implies they already have a project in mind, you have no idea how it’s going - if they’re stuck, alone, or not feeling it, they might be open to teaming up.

Chief of Staff (for business founder): Chiefs of Staff work closely with founders, see how companies are built from the inside, and often take on an immense amount of responsibility across ops, fundraising, hiring, and strategy. Many of them are itching to lead something themselves.

CTO as a service/Staff Engineer (for tech founder): ‘CTO’ alone is a mixed bag. Some CTOs are very hands-on, others haven’t touched code in years. Keywords like “CTO as a service”, “Fractional CTO”, or “Staff Engineer” tend to surface people who are still very much in build mode. They’ve got seniority and they’re used to being a one-person tech team - perfect for an early-stage cofounder role where they’ll be actively building from day one.

How to make sure they’re the right cofounder

There are plenty of great cofounder questionnaires out there and they’re all useful. But instead of listing every possible question, let’s zoom in on what actually matters most when assessing a potential cofounder.

At the core, you’re looking for alignment on the three key ingredients:

1. Complementary skills and personalities

2. Transparency, trust, and communication

3. Alignment on 5 non-negotiables values:
   - Long-term vision
   - Ambition
   - Commitment
   - Ways of working
   - Culture

The key questions to ask

Assessing complementary skills and personalities

You have to start by looking inward. Before you can evaluate someone else, you need to understand your own strengths and gaps.

Self assessment questions:

• What energizes you?

• What kind of dynamic drains you?

• Who brings out your best work and why?

• What kind of people or situations create friction for you?

Here are some questions you can ask when you’re ready to start assessing complementarity of skills and personalities:

Questions to ask potential cofounders to assess complementarity:

• What are you great at that you genuinely enjoy doing?

• What do you not want to be responsible for?

• If we started tomorrow, what part of the business would you naturally take ownership of?

• What’s one thing you’ve done in past roles that you’d want to avoid doing again?

• How do you recharge when things get intense?

Remember, you’re not just assessing hard skills, you’re also looking at perso- nality complementarity, which can be a bit trickier. It’s less about ticking boxes and more about trusting your gut. Try to gauge whether your personalities complement each other without clashing. One key way to avoid conflict down the line, even if your personalities are quite different, is to make sure you’re aligned on core values.

Assessing the five non-negotiable of value alignment

The harder layer to assess is value alignment. This is about alignment on the company you want to build and how exactly you are going to build it.

Because you’re not just starting a product together, you’re building a business, a culture, and a long-term partnership. If you’re misaligned on the fundamentals, it will be hard to get through it.

The secret to this step is to face the tough questions head on. It might feel awkward asking “are you comfortable working on weekends?”, but if you think it’s key and your cofounder has a strict no-weekend policy, you won’t have really gotten to the bottom of your values here. And it’ll cause problems down the line.

At the heart of value alignment are five non-negotiables:

Aligning on values as cofounders is foundational as those early shared beliefs often shape the company culture.

Assessing transparency, trust & communication

Throughout the cofounder matching process, one thing that matters just as much as the assessment of complementarity and values is how you communicate.

Transparency is a core ingredient in any successful cofounder relationship. If you’re not able to speak openly with each other now, during the early stages, it’s unlikely that will magically change later when the stakes are higher and the pressure is on.

That’s why one of the best signals to look for early on, and that means during the assessment part, is how naturally transparency flows between you. Are you both willing to be open?

Insight from Amaury Sepulchre, Partner and Cofounder of Hexa

The best way to build trust early is to go first. Be the one to open up. Here’s how to lean into that:

Talk beyond the business.

Get personal. Share your aspirations, fears, and dreams.

• What drives you beyond the startup?

• What scares you about this journey?

• Are there personal constraints (finances, family, health...) that might shape how you show up?

Be upfront about your weaknesses.

• You’re not perfect and odds are your cofounder won’t be either. So don’t
  be afraid of being vulnerable and talking about your weaknesses during the process, it will help your cofounder open up too and create a space of transparency.

The earlier you establish a foundation of openness, the easier it becomes to navigate the inevitable hard conversations ahead.

Don’t rush the process

You need enough signal to feel confident this is someone you can build a company with. One great conversation isn’t enough. Before committing to something this high-stakes, it’s worth spending a lot of time with your potential co-founder, in lots of different contexts.

Have at least three conversations

Chemistry in a meeting is one thing, but trust is built across different contexts. We recommend at least three conversations before making any kind of decision, ideally including one informal one (a walk, a coffee, a dinner). Three is the bare minimum - ideally, you’d have more. It helps you both drop the pitch and just talk like people.

Work together for a week

You’ll learn more from spending a week building together than from any number of conversations. Once you’ve had a few chats and things feel promising, this is often the phase where the rest of your questions get answered.

We usually recommend founders run a short POC together. You might spend that time mapping out what the product could be, exploring different directions, and talking to a few design partners. The output doesn’t matter as much as the process - how ideas evolve, how you handle ambiguity, and whether it feels like progress.

A week is usually enough to get a meaningful sense of the dynamic. If you’re still unsure after two, it’s probably worth pausing to reflect on why.

Get a second opinion

Ask someone you trust to give you an outside perspective, ideally someone
in the same domain as them. If it’s a technical profile, a friendly CTO might be willing to review their GitHub, or talk to them for half an hour to assess their technical skills. If it’s a business profile, ask a CEO or business lead to skim your notes about them, or ask if they’d mind having a quick phone call. Especially if they’re in your existing network, most people are usually happy to help!

Run references

A very important part of the process that should never be skipped. We recommend speaking to two managers and two people they’ve managed. And it should go both ways, invite them to do the same for you. You’re both making a serious bet, and references help you make it wisely.

Common founder mistakes

We’ve experienced a lot of founder matching processes at Hexa, and some of the same thinking traps show up again and again when people are choosing a cofounder.

Here are the four we see most often:

“We need to click instantly”: The first meeting is just the first step. If it feels a bit clunky, that’s normal - you’re still figuring each other out.

“We should be best friends”: You don’t need to be soulmates. You do need to enjoy working together, but it’s not about sharing secrets on a five-hour walk.

“I need to find someone like me”: This one’s sneaky as it often happens without realising. But sameness leads to blind spots and overlap, not balance.

“The more impressive the CV, the better”: Shiny résumés can be distracting. Grit, adaptability, and follow-through are harder to spot, but far more important.

Keep these in mind when evaluating potential partners. They come up more often than you’d think, and they usually point you in the wrong direction.

How to make it work

Define roles clearly

One of the most common reasons cofounders fall out is a lack of clarity about who’s driving what. You need to clearly define roles and ensure everyone is comfortable with their titles - whether CEO, COO, CPO, CTO, or others. This means making sure you’re aligned on:

• Who owns what (Product, GTM, Fundraising, Hiring, etc.)

• Who has final say in each area (and how you’ll make cross-cutting decisions)

Tip: Use a skills × passion matrix.

To figure out who should own what, list out all the key areas of the company. Then for each one, ask:

• Who’s strongest at this?

• Who’s most energised/passionate by this?

When assigning ownership, consider how the work energizes each person instead of relying solely on job titles or previous roles.

Splitting equity

There’s no one-size-fits-all when it comes to equity splits. Interestingly, when you look at SaaS companies at the IPO stage, most founding teams didn’t split equity equally.

This doesn’t mean unequal splits are always the right move, but it does suggest they can work under the right conditions. It also highlights that success depends on a range of factors.

That said, in our experience, when cofounders join at the same time and are equally committed, equal splits tend to work better. They reinforce mutual accountability and signal equal weight in shaping the company’s future. Unequal splits, on the other hand, can introduce subtle power dynamics, one founder may start to feel like an early hire rather than a true partner. Over time, that can erode trust, alignment, and motivation.

So if you can make equal work, we recommend doing it.

Bring in a cofounder coach early

Most founders wait until there’s already tension, but the best time to bring in a coach is when things seem to be going fine. That’s when you’re more open, less defensive, and most likely to build habits that stick.

A coach is there to help you talk about the things that are easy to ignore: how you make decisions, how you handle stress, what starts to build resentment when it’s not named.

They’ll help you spot the patterns early before they turn into problems. And more than anything, they create a space where you can be honest with each other, without it feeling personal.

Think of it like getting a trainer when you’re healthy, not after the injury. Prevention is the point.

Set up rituals

Rituals aren’t mandatory, but they’re something we actively encourage at Hexa. They help catch small misalignments before they grow and keep communication intentional.

That said, not every founder pair needs them. If you’ve already worked together for years, or if you’re side-by-side every day, those moments of sync might happen naturally.

Take Nicolas Benady, CEO & cofounder, and Nicolas Saison, CPO and cofounder, at Swan: they’ve known each other for 21 years and sit next to each other every day. Communication has become second nature, so a scheduled check-in doesn’t add much value. But for most cofounder pairs rituals can make a real difference.

Insight from Matthieu Vaxelaire, Partner at Hexa

A consistent weekly check-in can go a long way in the early days of a cofounder relationship. So can short retros after a big push, or async updates when you’re both moving fast in parallel. What matters is having regular, intentional moments to check in with each other.

One habit that has worked for many Hexa founders: a simple weekly self-assessment. Once a week, each cofounder shares:

• What went well this week

• What could’ve gone better

• One thing they’d like feedback on

• What keeps them up at night

It might feel a little formal at first, but over time, it builds a habit of alignment, which is what keeps relationships healthy under pressure.

Realign on values - not just once

Alignment on values isn’t a one-and-done conversation. It’s not enough to have a deep discussion during the cofounder “dating” phase and then assume you’re good for the rest of the journey. Priorities shift and what felt aligned six months ago may need revisiting today.

That’s why great founding teams treat value alignment as a recurring practice. They revisit the five non-negotiables — ambition, vision, commitment, ways of working, and cultural values — regularly, especially during key inflection points: fundraising, hiring, pivots, or even personal life changes.

Some founders go as far as setting a dedicated quarterly check-in, just to talk about alignment. Because in a cofounder relationship, misalignment doesn’t happen overnight. It creeps in slowly, silently, until one day you realize you’re running in different directions. Regular realignment keeps that from happening.

In 2016, two founders on opposite sides of the world had the same idea: a better way to take notes. One was working quietly on the first version of Notion. The other was building Slite. They didn’t know each other. They hadn’t copied one another. The truth is, the world was simply ready for that idea.

That’s the thing about startup ideas, they don’t usually come from lightning bolts or shower thoughts. They come from timing, noticing the shift in what’s possible, or seeing something broken or outdated, and realizing that now, finally, it can be done differently.

People often think of startup ideas as one-liners. But the best ones are more like frameworks that emerge from action, pattern recognition, and curiosity, not from waiting until we have a eureka moment.

In this chapter, we’ll show you what a startup idea really is, how to tell a weak idea from a strong one, and how to find ideas worth pursuing. You’ll learn the three types of startup ideas, the signals that matter, and the simple frameworks we use at Hexa to stress-test potential ventures before we write a single line of code.

Our goal is to help you go beyond “I have an idea” and move toward “I have a well-thought-out plan to build something exceptional.”

What’s a startup idea?

An idea is much more than a problem to solve

A lot of the startup advice out there simplifies building a startup into just identifying a problem and solving it. As a result, many founders spend their energy searching endlessly for problems to tackle. While this advice isn’t incorrect, it’s overly simplistic and limiting.

A robust startup idea is multidimensional. It must include:

• The problem → what specific pain point or need are you addressing?

• The solution → what product will you build to address this problem?

• The Go-To-Market strategy → how will you ensure your solution is known and used by customers? And how will it capture value

Perhaps part of the confusion arises from the word “idea” itself. It reminds us of a sudden flash of inspiration or a Eureka moment, something captured in a single catchy sentence. But a genuine startup idea is in fact more structured than this “aha” moment.

Rather than calling it a “startup idea”, it might be more accurate to describe it as a “startup framework.” This framework isn’t static, it’s something you continuously explore and refine through research, feedback, and real-world testing.

Every great startup innovates somewhere

At the heart of every successful startup lies innovation. Innovation doesn’t necessarily mean creating a groundbreaking technology or inventing something entirely new from scratch. Instead, innovation involves rethinking at least one key dimension of your idea:

• Problem: discovering and addressing a previously unnoticed or unaddressed problem.

• Solution: creating a totally new product that didn’t exist before or a signifi- cantly improved solution.

• GTM Strategy: developing an entirely new or more effective method to acquire customers or deliver your product to the market.

Excelling in just one area can be sufficient to create substantial value and differentiation in the market. For instance, innovating primarily in your GTM strategy can significantly expand your market reach, even if the problem and solution are already well-known.

Not all ideas are equal

You’ve probably heard the line: “It’s 99% execution, 1% idea.” It sounds good and practical but like most startup clichés, it’s only half true.

Even the most talented founders struggle to succeed if they start from a bad idea. Meanwhile, a well-timed, well-formed idea, even in the hands of an imperfect team can go surprisingly far.

Execution is the engine, but the idea is the map. If you start driving without one, you can end up working incredibly hard... in the wrong direction.

So before you pour in months of effort, take the time to ask the hard questions and validate your idea - using our framework that we’ll explore below. Not all ideas are worth building. The trick is knowing the difference.

How Stewart Butterfield the founder of Slack had a bad first idea

Stewart Butterfield’s journey is a really interesting example of how even exceptional founders can stumble initially due to a flawed idea. Butterfield, known today as the visionary behind Slack, initially invested his passion and resources into a gaming startup called Glitch, an online multiplayer game designed around creativity and exploration. Despite significant investment and effort, Glitch struggled to attract an audience and eventually shut down due to its failure to resonate broadly enough with users.

However, the end of Glitch became the unexpected beginning of something much bigger. In developing Glitch, Butterfield’s team had created a custom internal messaging platform to enhance communication and workflow among employees. Recognizing its potential, they pivoted from the game to refining this communication tool, ultimately launching Slack. Unlike Glitch, Slack quickly found product-market fit, dramatically reshaping workplace communication by significantly reducing dependence on emails and fostering collaborative team environments. The idea’s inherent strength transformed Slack into a global success, eventually leading to its acquisition by Salesforce for $27.7 billion.

Butterfield’s story illustrates the importance of starting with a strong idea, no matter how skilled the execution, success ultimately hinges on an idea that genuinely meets a need.

How to find an idea

The three categories of startup ideas

To effectively find startup ideas, it’s essential to first understand the three distinct categories they typically fall into. Throughout this section, we’ll give you our framework of the different types of ideas and provide actionable strategies for identifying ideas within each type.

There are three main categories of ideas:

Category 1: a problem-solution fit
Category 2: a market or technology shift
Category 3: a category creation

Category 1

• Source of the idea: often comes from a personal frustration, industry knowledge, or insight into a market.

• The problem: it exists today and is clearly recognized and felt by a specific group, and people are actively looking for a better solution. There may be no existing solution, or the available options are bad or outdated.

• Solution type: you create a new or better way to solve an already known issue.

• Example: Superhuman → born from personal frustration and deep user insight.
  Its founder, Rahul Vohra, observed that despite being a mission-critical tool, email had become slow and joyless, especially for power users like executives and founders. Rather than relying on a new technology break- through, Superhuman focused on obsessive UX, speed, and thoughtful features to reinvent the email experience. The problem was obvious; the innovation was in crafting a better solution for a well-understood need.

How to find them:

• Tap into personal frustrations: think of annoying, repetitive tasks you encounter regularly. Are there products or services you’ve used that feel outdated or inefficient? Chances are, if you’re frustrated, others are too.

• Leverage industry expertise: reflect on jobs you’ve held or industries you’re familiar with. What processes seem broken or overly complicated? Industry-specific insights often reveal problems that outsiders might not see.

• Engage with industry experts: VCs, investors, and founders see a high vo- lume of startup ideas and understand which problems are worth solving. Ask investors what types of companies they are excited about and what major pain points remain unsolved in their sectors.

• Observe what people are already trying to solve manually: anytime you see people building workarounds, copy-pasting between tools, hiring extra help, using Notion or Google Docs to track something, that’s often a sign that no good product exists yet.

• Examine existing solutions critically: look at the tools, software, or services available. What features are missing? Where do reviews or online forums show consistent dissatisfaction? These gaps are opportunities to offer something better.

Category 2

• Source of the idea: Comes from a new technology, market trend, or regu- latory change that makes solving an existing problem dramatically better, faster, or cheaper.

• Problem: it is already recognized and solved by existing solutions, but new technology enables dramatically better ways to solve it.

• Solution type: you rethink the solution using new capabilities that didn’t exist before.

• Example: Figma → Figma would not have been feasible five years earlier. Its core differentiation, real-time multiplayer editing in the browser, has been made possible only because of recent advances in web technologies like WebGL, WebAssembly, and CRDTs (conflict-free replicated data types). The “aha” wasn’t just that design tools needed collaboration; it was that now, for the first time, we could bring a Google Docs-like experience to a traditionally desktop-native category.

How to find them:

• Become an expert in the emerging technology of your choice: stay updated on advances in AI, web3, green tech, or other cutting-edge fields. When
  you understand what the tech actually makes possible, and its current limits, you’ll start to see which old problems can be solved in totally new ways.

• Track market and regulatory changes: New laws or regulations often unlock previously blocked opportunities. For example, GDPR created demand for better data privacy tools.

• Talk to users in legacy industries: Professionals in fields like law, logistics, finance, healthcare, or government often know their tools are outdated, but they may not know what’s now possible. Your role is to connect the dots between what’s painful and what’s newly possible.

• Map existing problems against new capabilities: Pick a known industry or workflow (e.g. expense reporting, insurance claims, knowledge manage- ment) and ask: How would this work if it were invented today, with the new tools we now have? This reframing often reveals huge efficiency gains or user experience upgrades that legacy players can’t easily deliver.

• Evaluate recent startups and failures: study what worked and what didn’t. Sometimes a startup fails because the market wasn’t ready. If conditions have changed, the idea might now be viable.

Category 3

• Source of the idea: a paradigm shift and dramatic change in human behavior that create opportunities for entirely new ways of doing things.

• Problem: it may not even be recognized by the market yet, as users haven’t identified it as an issue that needs solving. Often, these problems are deeply embedded in existing behaviors or accepted as “just the way things are” until an innovative solution reveals a better alternative.

• Solution type: instead of improving something that exists, the startup creates an entirely new way of doing things.

• Example: Airbnb → before Airbnb, staying in a stranger’s home wasn’t seen as a viable or safe alternative to hotels, and homeowners weren’t thinking
  of their spare rooms as income-generating assets. The problem wasn’t clearly defined but Airbnb uncovered a latent opportunity. The break- through wasn’t technological, it was behavioral. Airbnb created an entirely new way of thinking about lodging. It challenged cultural norms, built systems to create safety and reliability (reviews, identity verification, secure payments), and shifted what people considered acceptable when traveling, effectively creating the peer-to-peer hospitality category.

How to find them:

These are definitely the hardest ideas to find, but they also offer the greatest potential rewards.

• Study behavior, not just problems: ideas often emerge when you observe people doing things inefficiently, inconsistently, or with hacks, even if they don’t complain about it. Watch for friction people have simply accepted as “normal.”

• Explore cultural shifts and changing norms: big behavioral shifts often change how we do things and our mindsets. Pay attention to how younger generations communicate, travel, learn, work, and spend. What are they doing differently from the past, and what does that open up? The rise of the sharing economy came with shifting attitudes toward ownership , enabling platforms like Uber, Airbnb, and Depop.

• Question assumptions: challenge existing conventions. What if people didn’t have to own something to use it? What if processes that seem standard (e.g., commuting to an office) could be completely rethought?

The different types of startup ideas

Insight from Florent Quinti, Partner at Hexa

Today’s AI startups provide a perfect illustration of category 2 where a technology shift enables a better solution.

Today’s AI boom is a textbook example of a market or technology shift unlocking dramatically better solutions to known problems. The underlying problems - like customer support inefficiencies, content creation, document summarization, internal knowledge retrieval - have existed for years and were previously tackled through human labor or traditional software tools. With the arrival of large language models these same challenges can now be addressed faster, cheaper, and with far less manual intervention.

What makes this shift particularly fertile for startup ideas is how horizontal AI is: it touches nearly every function across industries. For example, the rise of AI agents in enterprise software reflects a rethinking of user interfaces and workflows, that used to be static forms can now be context-aware and thus personalized.

Founders who understand the underlying AI capabilities are in a unique position to revisit mature categories, like CRM, legal tech, finance, or education, and rebuild them from first principles, unlocking 10x improvements in speed, cost, or user experience.

Insight from Thibaud Elzière, Cofounder and Partner at Hexa

Ideas belong to their era, not to people
Great ideas are rarely unique. They tend to appear in multiple places at once because the conditions are ripe for them to emerge. In 2016, I had the idea for a simpler, more collaborative note-taking tool, that became Slite. That same week, I discovered a remarkably similar product on Product Hunt: an early version of Notion. It wasn’t plagiarism or bad luck. It was just that the world was ready for that idea.

That experience taught me something foundational: ideas belong to their time, not to their creators. They emerge at the intersection of cultural shifts, new technologies, and evolving user behaviors. The best entrepreneurs aren’t necessarily the first to think of something, they’re the ones who recognize when an idea’s moment has arrived and move fast to bring it to life.

The best ideas are born from action
You don’t find startup ideas by sitting around and waiting for inspiration. You find them by doing. Building companies, launching products, talking to users, testing tools, this is how you spot real problems and hidden opportunities. Every conversation, every experiment, every failure is a potential spark.

This is what I call building your ‘idea muscle’. People often think ideas come naturally to me, but it’s a habit I’ve built through years of practice. The more you engage with the world, especially from the inside, as a builder, the more patterns you start to see. You’ll notice inefficiencies, gaps, and signals that others miss because you’re close to the action.

And when you do stumble on something that excites you, even if it feels small, don’t let it fade. Talk about it, build something, test it quickly. Momentum matters more than perfection. An idea without action is just a passing thought.

Avoid rigid market analyses
One common trap is overanalyzing markets in a mechanical, top-down way. It feels productive to map out sectors, compare market sizes, and list competitors, but this often leads to obvious, uninspired, or already crowded ideas.

In my experience, good ideas don’t come from spreadsheets, they come from energy. From instinct. From something you saw, felt, or experienced that made you say, “there has to be a better way.” That’s why I prefer starting with the user, not the market.

Make sure it’s innovative

Once you’ve landed on a startup idea, your work isn’t done. Now it’s time to stress-test that idea using the framework introduced earlier:

A startup idea = Problem + Solution + Go-To-Market Strategy.

Every strong idea needs to innovate in at least one of these three areas. If it doesn’t, you may not have an idea worth pursuing, at least not yet.

This approach can also help you discover new ideas. Even if you find that your solution addresses an existing problem that already has solutions, you can still be innovative by developing a completely new go-to-market strategy.

Spotlight on Hexa’s portfolio

Several of our early companies, Mailjet, Mention, Aircall, Textmaster, and Front, emerged from problems Thibaud Elziere identified at Fotolia, where developers were spending too much time building internal tools. As Hexa expanded, we discovered new opportunities through challenges we faced while launching our first companies and running Hexa itself (leading to Swan, Spendesk, and Folk). Later, the emergence of web3 and AI inspired us to rei- magine software in new ways.

Examples from Category 1

Swan

When Upflow founders Alexandre Louisy and Barnaby Malet attempted to integrate third-party BaaS providers, they quickly ran into significant challenges.

That experience led us to build Swan, a developer-first BaaS platform that was modern, well-documented, and much faster to deploy. It wasn’t about inventing a totally new technology or inventing a category, it was about improving it, based on deep firsthand pain.

Folk

Folk started as an internal tool. We just wanted a better way to manage our contacts as a team, something between a CRM and a shared address book. Spreadsheets were clunky, CRMs too rigid. We built something small, just for us. But once we shared it, we realized how many other teams felt the same pain. That’s when we decided to turn it into a real product. The need was obvious once you had lived it, but almost invisible from the outside.

Examples from Category 2

Spendesk

Spendesk was born from a common pain point observed by both the Hexa team and the Spendesk founders: in some companies, employees were literally photocopying the company card to handle expenses. This small, easily overlooked behavior revealed a much deeper structural issue in how businesses manage spending. Around the same time, a new technology was emerging, the digital wallet, enabling companies to issue and control virtual or physical payment methods.

Instead of just building a better expense report tool, the Spendesk team flipped the model: what if every employee could access secure, trackable payment methods with built-in controls?

Aircall

The initial insight came from Thibaud Elziere’s time at Fotolia, where the
team relied on Skype to manage customer support across multiple countries. Skype’s ability to generate local numbers that redirected to an online application was clever, but it broke often and clearly wasn’t built for business. Still, it pointed to something bigger.

But the technology was shifting. Cloud infrastructure made it possible to build software without integrating with legacy telecom stacks. WebRTC, the technology behind Google Meet, was getting standardized and deployed widely, enabling any browser to become a phone app. And lastly, teams were moving toward more collaborative tooling, thanks to companies like Atlassian and Slack.

Put it all together, and the opportunity was obvious: build a fully cloud-based phone system, designed for modern teams from day one.

Tandem

In the past, customer success was human-led, or support was limited to static help docs and chatbots. Now, with embedded AI copilots, Tandem reimagines support as context-aware, real-time, and continuous, tailored for each user, inside each tool.

Without recent advances in LLMs and embedded AI, this type of seamless support layer would have been impossible. It’s a perfect example of taking a known problem (users struggling with software) and using emerging technology to solve it in a 10x better way.

Multis

Multis was a web3 company we launched in 2018, born from our desire to deeply understand an emerging technology by building within it. We spent months exploring the web3 ecosystem hands-on not with a fixed idea, but with the conviction that expertise comes from action. By immersing ourselves in the space, we uncovered a clear pain point: web3-native companies were completely underserved when it came to managing their finances. It’s a textbook Category 2 idea: the problem was known, but only with the rise of web3 infrastructure could a truly native solution emerge.

Tengo

Tengo uses AI to analyze tender opportunities, summarize requirements, and auto-fill key parts of the application. This was technically impossible until recently. With LLMs, Tengo turns dense legal and administrative documents into manageable workflows. A classic case of AI unlocking access to an old, painful process.

Examples from Category 3

Front

At first glance, Front seems like a classic Level 1 idea: solving the well-known pain of shared inbox chaos with a better, collaborative interface. But dig deeper, and you can make a strong case that Front actually helped create a new category - collaborative email - and nudged users into a different behavior paradigm. Traditionally, email was seen as a private tool. Teams that needed collaboration often resorted to hacks but rarely questioned the premise that email itself could be a shared workspace. Front redefined how teams could work through email together, with features like internal comments, assignments, and analytics baked into the inbox.

What makes this feel like Level 3 is that the problem wasn’t always explicitly stated by users. Many teams had accepted clunky workarounds as “just the way it is, and hadn’t yet imagined email could be truly collaborative. Front had to educate the market, change expectations around what an inbox could be, and essentially invent a new category: the shared collaborative inbox platform. Much like how Slack reframed internal communication away from email, Front reframed email itself. That’s classic Level 3: uncovering and solving a latent problem by introducing a better way of working that changes behavior.

How to validate an idea

Validating an idea isn’t a one-time checkpoint, it’s a continuous, iterative process. At Hexa, we’ve developed a simple but powerful three-step approach that’s helped us validate over 50 startup ideas, and invalidate many more. This framework doesn’t guarantee success, but it ensures we never build blindly.

Here’s the process we follow every time:

1. Become an expert
2. Scope the idea
3. Create wireframes

Become an expert

Even when you’re experienced in a sector, it’s tempting to jump straight into building. But relying solely on your own perspective is risky. To build so- mething truly useful, you need a deep, nuanced understanding of your users, their goals, workflows, frustrations, and context. And the only way to get that is by spending time with them.

This starts with a simple principle: genuine curiosity. Great validation doesn’t come from proving your assumptions, but comes from challenging them.

The best way to get close to the problem is to talk directly to the people living it. A good rule of thumb is to interview 25–30 people in your target audience, ideally through 30-minute calls. You’ll notice that once you’ve completed enough interviews, your learning curve starts to flatten. We call this the 80% rule: aim to learn 80% of what there is to know. Don’t get stuck chasing the remaining 20%, you’ll discover it as you build.

Here are a few best practices we’ve learned over time:

• Observe workflows, don’t just ask about them. Ask users to show you how they currently do things. Often, they can’t articulate pain points directly, but you’ll spot them in action.

• Ask short, open-ended questions. Your best tools are “Why?”, “How?”, and “Can you show me?”. Let users talk as much as possible - you’re here to listen, not pitch.

• Avoid leading questions. Stay neutral and open. If you’re fishing for valida- tion, you’ll get it, and it won’t be worth anything.

Scope the idea

Once you’ve developed a solid understanding of your users and the problem you’re solving, it’s time to get specific. That means moving from abstract insight to a clearly defined concept: what exactly are you building, for whom, and why now?

At Hexa, we’ve learned that writing things down is a critical part of validating an idea. It’s a way to force clarity. Scoping the idea early helps you confirm that:

• You’re solving a real, well-defined problem

• You understand who you’re solving it for

• The product you’re planning to build is a logical, focused response to that problem

• You’re not trying to do too much too soon

Here’s our template:

Summary
In one sentence, what are you building?

Problem
What is the problem you’re solving? Who’s experiencing this problem (persona? in what kind of companies)? What’s the frequency and the intensity of the problem?

Solution
What product will you need to build to solve the problem? You need two things here:

• A precise scope for the MVP

  • Including in the 2nd iteration mockups/wireframes and/or detailed functional definition of what you want to build

• A longer term vision of the product you want to build

Get into the details of your software (interfaces, workflows, inputs, outputs, interactions with the user, data structure...)

Implementation
You need here to be able to articulate how the product works for the user: how they will setup their software, what they will do with it, what is the input, the output...

Go-to-market
What kind of personas/ICP do we want to target? How are you going to find the first 100 users? What will the business model look like and how will we capture enough value?
In the longer term, how do you see GTM evolving?

Market size
To start with an idea, we don’t need a precise market sizing, but rather insight on how this can become a big market and how we can build a company larger than $100m ARR. What geographical areas do we want to target?

Opportunity
Why now? What have you understood that other people haven’t?

Competition
Brief list of players in the same space

Hypotheses
What do you need to believe in order for your startup to be successful?

Discovery
What do we need to test in discovery interviews to confirm the opportunity? Whom could we interview for this?

Create wireframes

Once the idea is scoped, the next step is to make it visual. Wireframes are one of the most powerful tools in early-stage validation, and one of the fastest ways to bring clarity to what you’re building. No styling, no branding,no pixel-perfect mockups, just clean, low-fidelity layouts that show what the product does, not how it looks.

At Hexa, we never skip this step. We always build a quick set of wireframes, typically 3 to 4 key screens, within the first few days of shaping an idea. It’s the fastest way to turn abstract thinking into something tangible, something you can point at and challenge together.

That said, wireframes at this stage aren’t always visual screens. For a vocal AI agent, they might be audio interactions or dialogue flows; for an API-first product like Swan, they might take the form of a first draft of API documentation.

More on how to build low-fidelity wireframes in chapter IV.

The timeline for idea validation

Insights from Quentin Nickmans, Cofounder and Partner at Hexa

Signs you shouldn’t validate your idea

• You’ve lost the spark. If the idea no longer excites you, don’t ignore that feeling. Building a startup is hard, if you’re not energized by the problem, it only gets harder.

• Customers don’t actually care. You’ve talked to potential users, and the feedback is lukewarm at best. They wouldn’t pay for the solution, or worse, they don’t even see the problem. That’s a vitamin, not a painkiller.

• The market is too small. Even if the product works, the upside just isn’t big enough to build a venture-scale company.

• VC interest is drying up. If funding in the space is shrinking, ask why. It could signal structural issues with the market or a technology that is about to be replaced. Which doesn’t signal anything positive.

Signs you shouldn’t give up on your idea

• A major competitor is launching. Don’t walk away just because a big competitor shows up. Back in 2014, we scrapped our plans to launch an eSignature company after a major player hit the market with a massive funding round and a bold promise: eSignature would be free for everyone. It shook us. But here’s the twist — that company had actually been building in stealth for 18 months and eventually launched... with a paid product. The space wasn’t closed. There was still plenty of room to innovate and win with a different vision. The lesson? Don’t get spooked by noise.

• Market experts tell you it’s a bad idea. Deep conviction matters more than expert approval. Look at Aircall. At the idea stage, Aircall called some early investors and experts in the Telco space to ask if the concept made sense. The answer? “You’re crazy, landlines are dead, the future is mobile.” Every expert said it was a bad idea. But they were wrong. Aircall found fast, strong product-market fit. Why? Because they stayed focused on the user problem and ignored the noise. Industry insiders often have blinders on, they see the risks, not the opportunity.

Hexa ideas that didn’t pass the validation stage

One of the fastest ways to understand what makes a great idea is to study the ones that didn’t make it. The ideas we didn’t pursue often reveal just as much as the ones that did.

FRANCHISEx

The pitch:

FranchiseX is the all-in-one platform for franchisors to manage and scale their franchisee networks. From onboarding new locations to improving perfor- mance across the network, our goal was simple: help franchisors grow, faster, smarter, and with less operational overhead.

Why we didn’t validate:

After speaking with customers and investors, we realized the market just wasn’t big enough to build a breakout company. The pain existed, but the upside was capped.

WEALTHx

The pitch:

WealthX is an AI-powered workspace for wealth management firms and advisors. By streamlining workflows and automating routine tasks, it helps advisors and assistants focus on higher-value activities, ultimately driving better client outcomes.

Why we didn’t validate:

Too few customers control too much of the market. Nearly half the revenue comes from just 50 firms, and the next 26% from another 500. That means missing one major player, especially in a country like France, means missing a meaningful chunk of ARR. On top of that, the product wasn’t enough of a wedge to become the operating system for wealth managers. It felt anecdotal next to core software.

FREIGHTx

The pitch:

Carriers, brokers, and 3PLs waste hours on manual tasks, quoting, order processing, customer service, ops management. These inefficiencies kill margins and limit growth. AI agents purpose-built to automate these workflows could dramatically improve efficiency, service quality, and scalability.

Why we didn’t validate:

This wasn’t the right market to launch in. Margins are thin, and the customer base is either huge enterprises or deeply entrenched legacy SMEs, both hard to access and not feeling acute productivity pain.

More importantly, Europe lacks the kind of fragmented, independent freight brokers that exist in the U.S. Freight in Europe almost always crosses borders, making it a game of corridor-based freight forwarders rather than small brokers managing hundreds of carriers. Without those small players, the wedge disappears.

To get the Hexa Roundup delivered every month to your inbox, don’t forget to subscribe below - or go to your account settings to turn of Hexa Roundup emails.

Subscribe now

📖 Chapter 1 of The 10x Method: becoming a 10x founder

The first chapter of Hexa’s book is now available online. It covers the 10 traits Hexa has identified across 100+ founders, from balancing conviction with adaptability, to keeping your ego out of decisions, to thinking global from day one. Each trait comes with concrete advice and quotes from founders like Mathilde Collin (Front), Olivier Pailhes (Aircall), and Christophe Pasquier (Slite). The core idea: the DNA of your company starts with you, and these are skills you can build.

→ Read chapter 1

🧗 How to get to €25M ARR without raising a euro

Hexa Start Partner Mat Vaxelaire sits down with Émeric Ernoult, cofounder of Agorapulse, now the #1 social media management tool in France, built entirely bootstrapped. The conversation covers 10 years of evenings and weekends before the product even existed, setting an ultimatum of 1,000 customers or quitting, rebuilding the entire tech stack mid-growth, and a cofounder relationship that required actual therapy to survive. Émeric is honest about all of it, including the one hiring mistake he repeated for 15 years.

→ Watch the episode

🏞️ Why the era of vertical SaaS founders is now

Horizontal SaaS is showing its limits. Salesforce now has twelve “Industry Clouds,” essentially admitting that one-size-fits-all doesn’t work. Meanwhile, vertical players like Toast and Mews are posting real profitability by embedding payments and lending directly into their products. In this article, Hexa Scale Partner Augustin lays out why domain expertise has become the real moat, how fintech can multiply revenue per customer, and why the “too small” markets are now worth billions.

→ Read the full piece

🔌 The CTO dominating the MCP market before it explodes

Frédéric Barthelet (cofounder of Alpic) talks with Louis Pinsard (cofounder of Dialog) about raising $6M to build the hosting and tooling layer for MCP, the protocol that lets AI models interact with external software. Less than a year after MCP’s creation, Alpic is positioning itself as the Vercel of this new ecosystem, complete with an open-source framework and a go-to-market that targets coding agents as much as developers.

→ Watch the episode

Portfolio news

🧱 Enobase joins Hexa Sprint to rethink the ERP from scratch

Welcome to Enobase, the newest company in Hexa Sprint’s AI for Work vertical led by Thibaud Elzière. Basile Cassiers and Gauthier Rodaro founded Enobase after seeing the same pattern everywhere: operations held together by spreadsheets and workarounds that were never meant to last. Their bet is a platform that adapts to how a business actually runs, rather than forcing it into a rigid system. Learn more.

⚖️ Plato joins Hexa Start to build an AI copilot for litigation

Ben Lasseri and Alexandre de Boutray cofounded Plato alongside Partner Mat Vaxelaire to tackle the repetitive work that eats up litigators’ time - recalculating inflation, sorting medical documents, drafting claims from scratch. The goal: let lawyers focus on strategy and defending clients. Already 1,500 cases created, one of France’s largest insurers onboard, and dozens of lawyers using it daily. Learn more.

⚡ Revolty joins Hexa Sprint to rethink home energy storage

Revolty is the latest company to join Hexa Sprint’s Climate cohort, spotted by partner Maxime Brousse. Founded by Mathilde Janicot and Valentin Heuzé, the company repurposes Europe’s second-life lithium batteries into high-performance residential storage, closing the gap that keeps solar households from capturing the full value of the energy they produce. Already €2M in secured orders, with certification planned for 2026. Learn more.

🌹 Rose launches from Hexa’s SF House to rebuild inbound

The HubSpot-era inbound playbook no longer matches how B2B buyers actually behave. Johary Randria and Benoit Pothier cofounded Rose with Hexa Start partner Florent Quinti to fix that: a suite of AI agents that engage, nurture, and convert website visitors autonomously. Already 10,000+ conversations handled, up to 6% engagement rates, +20% conversion uplift, and companies like Pennylane, Skello, and AB Tasty onboard. Rose is also the first startup out of Hexa’s San Francisco batch. Learn more.

Hexa news

🚀 Raphaël Vullierme joins Hexa Start as Partner

Raphaël Vullierme — founder of Luko, repeat entrepreneur, Accel Scout — is joining Hexa to co-found and build at the frontier of AI and Defense. His conviction is that the most important customer of the next decade won’t be a human, but an AI. He’s already working from Hexa’s San Francisco House, pushing Hexa Start into what he calls “Ambient AI”, where technology becomes an active agent in the economy, not just a tool. Learn more.

🔎 Camille Goni joins Hexa as VP Talent & Ecosystem

Every year, thousands of entrepreneurs apply to Hexa, but only a handful build with us. Camille Goni joins to make sure the best ones keep coming. Previously leading the Talent team at Shine, her mission is to make Hexa the obvious choice for the most ambitious founders looking to build AI-native companies, and create the environment that multiplies their ambition. Learn more.

Jobs

• Talent Acquisition Partner at Hexa

• Founding Full-stack Engineer at Waniwani (HX26)

• Head of Revenue/late cofounder at Filiz (HX25)

• Founding Engineer at Panora (HX25)

• Founding Engineer at Plato (HX26)

• Legal Product Manager at Plato (HX26)

Thanks for reading! Follow us on Linkedin and Instagram to stay up to date between issues.

Cheers,

The Hexa Team

For this episode of PMF Collector (ex Start Mode), Matthieu Vaxelaire sits down with Émeric Ernoult, co-founder & CEO of Agora Pulse - the #1 social media management tool in France, and the 5th largest worldwide, built without raising a single euro.

It took 15 years. A failed first company. Two men working evenings and weekends for a decade. A co-founder relationship that survived 24 years, therapy included.

But what makes this conversation worth your time isn’t the destination. It’s the honesty about how ugly the road actually was.

You can watch the full conversation above, or read the key themes below.

1) The co-founder you didn’t choose — and why it worked anyway

Émeric didn’t meet Benoît at a startup event or through a warm intro.

They crossed paths at a student house in New York in 1993, during their first internship in the US. Years of reconnections followed — Paris, Washington, chance encounters through Benoît’s twin brother — until 1999, when both were back in Paris, restless, and caught up in the internet bubble.

“We made three PowerPoint slides and thought we’d become multi-millionaires.”

They didn’t. But they stayed together for 24 years — longer than Émeric’s marriage.

2) Ten years of evenings and weekends — and why he doesn’t recommend it

From 2000 to 2010, both founders kept their day jobs.

Émeric worked at a law firm. Benoît at a web agency. Their startup — Affinity, a white-label community platform — ran on whatever hours were left. Morning commutes, Saturday afternoons, Sunday evenings.

At peak, the business was doing €150K in annual revenue. After a decade.

“It should have died. We should have moved on. It wasn’t normal what we did.”

He doesn’t romanticize it. The model was broken — every client needed custom development, it was agency work dressed up as a product, and it never scaled. But something kept them going: the belief that the next pivot would be the one.

3) The insight that created Agora Pulse: sell what people use every day

In 2010, they stumbled into Facebook contest apps by accident — a campaign for the FIA featuring Sébastien Loeb went viral, and suddenly brands were calling.

They sold bespoke Facebook contests at €15,000 each. Within a year, fifty agencies were doing the same thing in Paris, and prices collapsed to €6,000. Then lower.

Meanwhile, they noticed a US startup called Wildfire productizing the exact thing they were doing manually — turning it into a self-serve platform. Google later acquired it for $450M and shut it down within a year.

The real question wasn’t contests. It was: what do community managers actually use every day?

“We asked ourselves: who are our buyers, and what do they need constantly — not once a year?”

The answer was a publishing and engagement tool. And that became Agora Pulse.

4) The ultimatum: 1,000 customers at €49/month — or we shut it down

By late 2011, Émeric was done with uncertainty.

He set a hard target: 1,000 paying customers at €49/month within one year. If they missed it, he was walking away. He’d spent too long, earned too little, and couldn’t keep convincing himself the breakthrough was coming.

It took three years, not one. But every single month — except December and August — the number went up. Three new customers. Four new customers. Slow, but always forward.

The thing that kept him going was a talk by Gail Goodman, founder of Constant Contact, called The Long, Slow SaaS Ramp of Death. Her message: the only question that matters is whether you’re moving in the right direction. Not how fast.

“Every month it progressed. Two months a year it went backwards. That was enough to hold on.”

5) Bootstrapped to €23M ARR — and what that actually costs

Agora Pulse never raised a meaningful round. A €250K grant from a public fund. That’s it.

By 2015 they crossed €1M ARR and reached profitability. By 2016, they grew 150%. By 2019, 90%. Today the company sits around €23M ARR, with 160 employees across France, the US, and internationally.

But Émeric is clear: the growth wasn’t linear. 2017 was nearly flat — Benoît spent the year rebuilding the entire tech stack from scratch, and zero new features shipped. When you stop shipping, customers stop growing.

“Once you’re profitable, the emotional rollercoaster is less violent. But you never feel like you’ve arrived. That’s the nature of the job.”

6) Building on someone else’s land — the risk nobody talks about

The existential vulnerability of a social media tool is obvious once you say it out loud: everything depends on Meta, Twitter, LinkedIn keeping their APIs open.

Émeric has lived through accidental outages that froze €23M in ARR for days at a time, with no clear escalation path, waiting for an engineer in India to push the right button.

“My next company will not be built in someone else’s garden. I know that for certain.”

The diversification across multiple platforms helps — losing one network doesn’t kill the business. But the dependence on Meta in particular remains a structural risk he carries every day.

7) Twenty-four years with one co-founder — the honest version

Émeric and Benoît never did the thing advisors tell you to do: write down what each of you expects, read it back, sign at the bottom.

They fought constantly. They went through stretches of barely speaking. At one point, they hired a coach — couples therapy, essentially — to get back to a functional working relationship.

What saved them wasn’t process. It was a deliberate decision Émeric made about how to look at his partner.

“I decided to look at him as the person without whom none of this exists. Full stop. The rest doesn’t matter.”

His advice to founders pairing up today: write down your expectations before you start, revisit them every quarter, and give each other actual feedback. He admits he never did any of it properly. That’s exactly why he knows it matters.

8) The one thing he’d fix if he started over

Émeric Ernoult’s answer comes from the book he’s writing.

The biggest mistake he made — consistently, across fifteen years of hiring — was bringing people on without ever defining precisely what success looked like for them. Not a vague job description. Not a list of KPIs. A specific, co-authored document that the person reads, agrees to, and can be held accountable against.

He calls it a What Does Success Look Like document.

“We spend so much time hoping people will figure it out on their own. They don’t. Because we never told them what ‘good’ actually means in their role.”

He’s been running this at Agora Pulse for eight years. He believes it’s the single biggest lever most founders leave untouched.

After working with over 100 founders, we’ve seen it time and again. The founders’ beliefs, behaviors, and decisions shape everything: the culture, the pace, the product, the team. The DNA of the company is set by the person at the top.

Building a 10x company requires becoming a 10x founder. That’s why The 10x Method starts with you.

We’ve identified the 10 traits we believe sets them apart. These traits are mostly learned through experience, so, if you don’t see yourself in every one of them yet, that’s not a problem. These are skills you can build.

As you read through them, ask yourself: How can I apply this today? Where am I already strong? Where do I need to grow? And what’s one thing I can start doing differently now?

1. Have a long-term vision, and the first step to get there

You need to be crystal clear about your vision. Where do you want your company to be in five years? What impact are you trying to create? How will your product shift behavior or redefine an industry?

That vision is your compass, but that’s the easy part in a way. What’s harder and where founders often go off track is the very first step. What you execute on and what needs to happen this week to move in that direction.

A helpful way to visualize this is as a domino effect: today is the first domino, next week is the second, and each one builds momentum toward the final domino, your long-term vision.

Each action today sets up the next until your long-term vision is achieved.

2. Balance conviction with adaptability

To build something new, you have to believe in something that doesn’t exist yet and convince others to believe in it too. That requires bold conviction in your long-term vision. You need to be all-in, willing to push through skepticism.

In parallel, you have to be radically open to feedback: your first instincts might be off, your messaging might not land, your target customer might not care or might not be who you thought they were. If you treat every signal as a threat to your idea, you’ll miss the truth. And if you ignore feedback entirely, you’ll keep refining something that nobody wants.

The solution is to simultaneously be convinced of your vision, and adaptable on the first step - and the second, and the third. Think of each early move as a hypothesis. Your job isn’t to defend it, your job is to test it. If it doesn’t move you closer to the vision, throw it out and try another.

3. Execute relentlessly

In the early stages (and often well beyond) founders must stay hands-on and deeply involved. What makes startups powerful is that the people with the vision are also the ones closest to the work. Unlike in big companies, where layers of management create distance, startups thrive because their founders are in the trenches: building, coding, selling, and learning alongside their users.

It’s about showing up every day to execute relentlessly.

4. Create velocity with routines

A great way to drive relentless execution and move fast is to build a strong cadence around it. At Hexa, we’ve launched over 50 companies using the same operating rhythm from day one: quarterly goals to set direction, and weekly sprints to maintain momentum.

Quarterly goals

Every 90 days, you zoom out to realign. A company-wide meeting is held to reflect on what’s been accomplished and to set clear goals for the next quarter. Each key area (product, go-to-market...) should leave with a focused set of OKRs. At the early stage, every quarter can bring defining moments: launching your MVP, signing your first users, or closing a funding round.

Weekly sprints

Once the quarter is set, everything comes down to the week. Great founders operate on a weekly cadence tracking core metrics and setting clear goals for the week. This steady rhythm is what builds real momentum.

5. Keep your ego out of it

Low-ego founders build stronger companies because they focus on what works, rather than on being right. They hire people smarter than themselves, admit when they’re wrong, and ask for help without hesitation. When something isn’t working, they pivot instead of clinging to an idea to protect their identity.

6. Build together

No founder builds a great company alone. The most ambitious goals are only reached when founders surround themselves with exceptional people who challenge them and complement their strengths. 10x founders stay open to feedback, listen to their employees, and create a culture of collaboration.

7. Think global from day one

Your home market is not your ceiling. In fact, it’s often your biggest constraint.

European founders in particular fall into the trap of building locally first and expanding later. But by then, it’s often too late, the product is overfitted to one market, the messaging doesn’t translate, and the culture isn’t built for scale.

10x founders flip this. They build international from day one:

• They speak English internally.

• They hire globally, even if the team is small.

• They design products with cross-border scalability in mind.

• They seek out international design partners, not just local champions.

You don’t need a global footprint on Day 1, but you do need global intent.

8. Be a missionary, not a mercenary

The founders who endure, and win, are chasing impact, not exits.

This doesn’t mean you need to start a nonprofit. It means you have a real belief in what you’re building, and a commitment to long-term value creation. Investors can feel that. So can your team. So can your users.

Mercenaries make short-term bets. Missionaries build long-term momentum.

9. Build resilience through purpose

What keeps you going when it gets really hard, when the deal falls through, when the product breaks, when the money’s almost gone?

It’s purpose.

The most resilient founders are deeply connected to their “why.” They feel their vision deeply. It’s personal. It fuels them when things get ugly. And that’s what allows them to keep showing up with energy and clarity when others would quit.

Resilience isn’t about pushing through everything blindly. It’s about having a reason that makes the struggle worth it.

10. Lead with kindness

The truth is, people want to help kind founders. Employees stay longer. Investors go the extra mile. Fellow entrepreneurs open doors. You’ll move 10x faster with a network that believes in you, and that starts with how you treat people.

This doesn’t mean being soft. It means being respectful, clear, generous with credit, and honest in feedback.

We published our book, The 10x Method, last Autumn.

Built on the back of 50+ startups, including Spendesk, Front, Aircall, Yousign and Swan, it's the repeatable methodology behind three unicorns and other breakout companies.

It was a victim of its own success, selling out within the first few weeks.

Lots of people have since asked us when it will be back in stock. Unfortunately, we won’t be printing any more physical copies…

Fortunately, we decided to go one better!

Starting today, we’re publishing The 10x Method in full, one chapter at a time, here on Hexa Media.

No paywalls, no waiting around for your copy to arrive, just the full methodology immediately at your fingertips.

For the next couple of months, a new chapter will arrive in the inbox of our subscribers every week.

We hope it helps more founders to beat the odds. Enjoy.

The 10x Method - Foreword

Let’s start with the obvious: most startups fail.

Only 1 in 10 seed-funded startups make it to Series A. The road is unforgiving, filled with moments that push even the most driven founders to the edge.

No one becomes an entrepreneur just to get rich. At least not the ones who make it. There are far easier ways to make money. Becoming a founder is something else. It’s a calling. It’s the relentless belief that the world can be better, and the irrational conviction that you’re the one who can make it happen.

Founders take this path for different reasons: to solve a problem no one else can, to build something that’s never existed, to create impact at scale. But they all share one trait: they are builders. They go from zero to one, from nothing to something, from idea to company.

Creation demands resilience most people don’t have and maybe shouldn’t. It requires optimism that borders on delusion. But it is also a highly rewarding adventure. There’s real magic in bringing something to life and watching it improve the world.

And because that journey is so meaningful, we want more people to take it. That’s why we’re sharing our method, 14 years in the making, built and refined at Hexa to turn bold visions into exceptional companies.

This is the method we’ve used to launch over 50 B2B software startups. It walks you through every step of the early stages of building a company: from finding the right idea and co-founder to building your MVP, acquiring initial users, and raising funds.

This book focuses on the early stage, before any Series A round is raised, because that’s where we’ve built deep expertise. We know what it takes to go from zero to one, because we’ve done it over fifty times at Hexa. These early decisions, what you build, who you build it with, how you launch, shape everything that follows.

But there are no guarantees. Every company is different and every journey is unpredictable. There is no magic bullet.

What exists, and what we’ve seen work, is a set of routines, beliefs, and ways of operating that dramatically increase your chances. They’re not hacks nor one-size- fits-all templates but methods that help you do the right things, in the right order, at the highest possible standard.

That’s what the 10x Method is.

It replaces scattered startup advice with a unified system rooted in experience. And it has helped turn early-stage ideas into globally significant companies.

We’ve seen it increase the odds of success tenfold: creating companies that go global, products that change behavior, and teams that scale from two founders in a room to hundreds of people across continents.

We’re talking about startups that leave a dent in the world.

The 10x Method is designed to build those kinds of companies. Ones with a real shot at becoming unicorns.

This book is for you if:

• You want to build or are currently building a B2B software company

• You’re at the early stage, maybe pre-product, maybe pre-idea

• You’re aiming to build something massive

• You plan to raise capital, not bootstrap*

• You’re driven by global impact, not local reach

*Bootstrapping is a perfectly valid path for entrepreneurs - not better or worse, just different. This book is written for founders who choose not to bootstrap and are building with external funding.

How to read this book:

The book is structured to follow the natural timeline of early-stage startup building (from the first spark of an idea to fundraising). Each chapter stands on its own and isn’t ordered chronologically necessarily - you might fundraise before you make your first hire! You don’t need to read it cover to cover to get value. Whether you’re validating an idea, hiring your first team member, or refining your go-to-market strategy, feel free to jump straight to the chapter that’s most relevant to you.

Even with the 10x Method, some Hexa startups haven’t made it. But the ones that did didn’t get there by chance. They got there by committing to excellence early and operating at a higher standard from day zero.

If you’re building a B2B software company and want to increase your odds of success, this book is for you.

We believe entrepreneurship drives progress. We believe the world needs more founders. And we believe that now is the time to build.

Every month, Hexa hosts an in-person fireside chat where a CTO shares their journey, learnings, and vision of the role.

Before the event, we film a short conversation to share their insights with those who can’t join us in person.

In this new episode of CTO Voices, Louis Pinsard (Cofounder and CTO of Dialog) sits down with Frédéric Barthelet (Cofounder and CTO of Alpic) the MCP infrastructure platform for deploying AI agents at scale.

Frédéric’s path led him to co-found Alpic in the summer of 2025 alongside four partners, raising $6M to build the infrastructure layer for a brand new ecosystem: MCP servers and GPT/Claude applications. In less than a year, he’s been navigating one of the fastest-moving spaces in tech - evangelizing a protocol that didn’t exist before November 2024.

What follows is a conversation about building in an emerging ecosystem, reaching developers through AI agents, and betting on a paradigm shift that feels a lot like the early App Store.

What is MCP and why does it matter?

MCP (Model Context Protocol) was born in November 2024 as a standard for communication between AI models and traditional software. It lets models retrieve context, trigger actions, and interact with the outside world.

More recently, the protocol was extended with the concept of apps, which add the ability to render UI directly inside a conversation with Claude or ChatGPT, and to surface discoverable services to users.

Frédéric’s bet: this is the new distribution channel. Just as the App Store created a new surface for products in 2008, MCP apps are creating a new surface inside conversational AI clients.

Alpic: Vercel for the MCP ecosystem

Alpic is to MCP what Vercel is to web apps, an infrastructure platform purpose-built for this new type of workload.

That means: hosting, scaling, monitoring, analytics, and authentication (OAuth, which is native to the MCP protocol) - all adapted for MCP servers and apps.

More recently, Alpic launched Skybridge, an open-source full-stack TypeScript framework for building MCP apps for ChatGPT and Claude. It comes with emulators, dev tools, and examples so developers can iterate fast without needing live access to GPT or Claude during development.

A GTM built for the age of coding agents

Reaching developers today isn’t what it was 10 years ago. The strong framework loyalty of the past (where mastering a tool gave you a real edge) has faded. Coding agents now help developers onboard into new technologies faster than ever.

So Alpic’s go-to-market reflects this shift. Rather than only convincing human developers, they’re also trying to convince the coding agents themselves.

Their approach: a Claude skill that teaches Claude how to build MCP apps, and recommends Skybridge as the best way to do it. A developer who asks Claude to build a GPT app will naturally end up using Skybridge, and from there, deploying on Alpic.

Evaluating AI-native distribution

Alpic has built an internal evaluation pipeline to measure whether their skill actually steers Claude toward the right outputs - not deterministic unit tests, but statistical scoring to track whether documentation and instruction changes move the needle positively or negatively.

The skill doesn’t just teach Skybridge. It also teaches Claude how to verify the output runs correctly, how to simulate widget rendering, and how to deploy, a full loop from build to production.

The App Store moment

Frédéric keeps coming back to one analogy: the first year of the App Store, the top-grossing app was a virtual beer that you could “drink” by tilting your phone. Nobody would have predicted it. Nobody could have imagined it before the ecosystem existed.

That’s where we are today with MCP apps. A new triptyque (user, model, and software) creates interaction patterns that weren’t possible before. The job right now is to help people imagine what they couldn’t imagine before.

Thanks for reading!

👉 Don’t miss our upcoming CTO Voices events - in-person gatherings for CTOs to connect, exchange, and share hard-won lessons.

📅 See all upcoming dates on luma.com/hexa

Salesforce stock dropped 43% in 12 months. Meanwhile, Toast—the restaurant software company—posted its first year of GAAP profitability in 2024, with 80% of revenue now coming from embedded payments and lending. Mews, which makes hotel management software, processed $10 billion in payments last year and reached $2.5 billion in valuation.

If you’re building vertical SaaS, these data points should matter to you, because they represent a fundamental shift in how software creates value. For the last 15 years, the venture capital playbook has been clear: horizontal beats vertical, the biggest TAM wins.

That playbook is breaking down, and the vertical SaaS founders I work with at Hexa Scale are starting to see it in their own businesses.

I think we’re at an inflection point, and most vertical SaaS founders don’t realize how strong their position actually is. You’ve been told your market is too small, your niche is too specific, your business isn’t venture-scale. Meanwhile, Salesforce has launched twelve “Industry Clouds” trying to become a vertical solution for healthcare, education, manufacturing, and every other sector. They’re spending billions trying to build what you already have: deep domain expertise and workflows designed for a specific customer.

The numbers are showing up in retention rates, in revenue per customer, in the economics of customer acquisition. Vertical SaaS companies consistently show higher retention than horizontal platforms, and the best ones are now layering in embedded financial services that can significantly increase revenue per customer. AI is strengthening their moat while simultaneously eroding horizontal SaaS’s.

Indeed, the underdogs are winning.

Horizontal SaaS is running out of steam

Horizontal SaaS has been the obvious startup bet for some time. CRM for everyone. Marketing automation for everyone. Project management for everyone. Build one product, sell it to every industry, achieve infinite scale - the pitch was elegant.

The model has started showing its limitations. HubSpot grew revenue by 18% in 2025, which sounds good until you realize that revenue per customer only grew by 3%. They’re adding customers, but those customers aren’t getting more valuable. The growth is coming from volume, not depth, and volume is an expensive game when your customer acquisition costs keep climbing.

The Salesforce story is even more revealing. They’ve now launched twelve “Industry Clouds”—for healthcare, education, financial services, nonprofits, manufacturing, government, retail, and on. Each one is essentially an admission that horizontal software doesn’t actually work horizontally. What they’re building now are vertical solutions bolted onto a horizontal platform, except the platform was never designed for this. It’s become a kind of digital Frankenstein, extended by armies of consultants who use it as a blank canvas because the out-of-the-box version doesn’t do what anyone actually needs.

For every dollar Salesforce makes, their partners make $6.19—the consulting ecosystem is five times larger than the product itself. Ninety percent of implementations require custom development, users only utilize half the available features, and 42% still store critical data in spreadsheets. Forrester reports that 70% of implementations fail to meet their stated objectives.

This is what happens when you try to make one tool serve every possible use case. You end up with something so general that it’s useful to almost no one without significant customization, and customization at enterprise scale costs a fortune.

Of course, horizontal SaaS isn’t going to collapse overnight. Enterprise software moves slowly, switching costs are brutal, and most large companies haven’t even finished moving off on-premise systems yet.

But new market entrants aren’t choosing these platforms the way they used to. Small and mid-sized businesses are adopting faster, and they’re increasingly choosing tools that were built specifically for them rather than tools that were built for everyone and no one.

Domain expertise = the real moat

While horizontal SaaS companies were raising massive rounds and chasing infinite TAMs, vertical SaaS founders were building something different. Software for labs. Higher ed administration systems. Tools for small construction firms, hotels, dental practices. The markets looked small—$50 million to $500 million addressable, not the $10 billion-plus that VCs wanted to see. The industries felt unsexy. The founders were often industry insiders, not product people from Google or Facebook, which meant the early versions had dated UX and clunky interfaces that looked “old school” compared to the polished horizontal tools.

VCs passed. Growth investors passed. The conventional wisdom was that these businesses would never scale, and the founders would never get the outcomes they deserved.

Meanwhile, these companies were building something that horizontal platforms couldn’t replicate. They understood the workflow at a level that no general-purpose tool could match. A dental practice management system built by someone who’s run a dental practice knows things that Salesforce will never know—how insurance verification actually works, how to handle operatory scheduling conflicts, how to manage the specific compliance requirements for patient records.

The retention numbers reflect this. Vertical SaaS companies see 35-60% higher retention than horizontal platforms because the software actually solves the problem it was designed to solve, without requiring an army of consultants to make it work. The CAC is lower because word spreads through industry networks—trade shows, professional associations, peer recommendations.

Many of these companies were bootstrapped or grew with minimal outside capital, which meant the founders kept control and built sustainable unit economics from day one. There were no insane burn rates, no growth-at-all-costs mandates from a board that needed a 10x return in five years.

3 forces accelerating the shift to Vertical

The landscape has recently shifted in three ways that particularly benefit vertical SaaS, and understanding these shifts is critical if you want to capitalize on this moment.

1. AI can’t replace knowing what to build

AI is very good at solving general problems. This is bad news for horizontal SaaS, because horizontal SaaS solves general problems. If your value proposition is “we’re a CRM for everyone,” and AI can help a company build a lightweight custom CRM in a few weeks, your moat disappears quickly.

Vertical SaaS has a natural defense against this: deep domain expertise. AI can help someone build a generic task management system, it can tell you what high-level workflows a hospital lab needs for specimen tracking, but not at the level of field experience, that gives you all the small details and edge cases that make a good product good. That knowledge still has to come from somewhere, and “somewhere” is either years of industry experience or a product that was built by people who have that experience.

Software creation costs are decreasing, which means the value is shifting to knowing what to build. Product design informed by deep domain knowledge becomes the differentiator, not the ability to write code. If you’re a vertical SaaS founder, this is working in your favor.

2. Fintech can 3-5x Vertical’s Revenue Per Customer

Traditional SaaS charges subscription fees: $50, $100, $500 per user per month. But that model is getting commoditized, and AI is accelerating the compression. It’s getting harder to charge meaningful subscription fees for software alone.

Meanwhile, fintech has become embeddable. Payments, lending, insurance, bank accounts—all of it can now be built directly into your product. And if you’re vertical SaaS, you’re already sitting in the middle of the financial workflow. Restaurant software sees every transaction, hotel software processes every booking, legal practice software handles every client payment.

Toast already makes 80% of its revenue from financial services, not software subscriptions. They lend $1,000 to $300,000 to restaurants through Toast Capital, underwritten on the transaction data flowing through their system. That lending business helped them reach GAAP profitability in 2024 after years of losses.

Mews processed $19.7 billion in payments in 2025 —double the prior year— and their valuation jumped from $1.2 billion to $2.5 billion. Clio, the legal practice management platform, is now over $300 million in ARR with billions processed annually through Clio Payments.

Stripe reports that 60% of all small businesses in America use a vertical SaaS platform, and over 15,000 vertical SaaS companies have embedded Stripe payments. The reason is simple: adding fintech can increase revenue per customer by 2-5x, and the unit economics are often better than subscriptions. You’re taking a small percentage of large transaction volumes instead of charging a fixed monthly fee.

If you’re a vertical SaaS founder and you haven’t thought seriously about embedded fintech, this should be a top strategic priority. It’s a completely different business model that compounds the advantages you already have.

3. Your customers will buy, not build

Your customers are not going to build their own software, even as AI makes it easier to do so.

Restaurants, schools, museums, labs, and dental practices are businesses at the end of the digitalization curve. We’re still talking about software adoption, about replacing pen-and-paper processes or on-premises legacy tools with digital workflows. These customers don’t have internal tech teams. It’s not their job to code, and even with AI lowering the barrier to software creation, they’re not going to start.

This is different from the enterprise customers that horizontal SaaS serves. Large companies have engineering teams that can evaluate build-versus-buy decisions. When AI makes it easier to build custom internal tools, those companies will increasingly choose to build. But your customers—the restaurant owner, the dental practice manager, the hotel operator—they just want a tool that works. They will buy, not build.

This is a structural advantage that’s easy to underestimate. The threat that AI poses to horizontal SaaS—that customers will build their own solutions—doesn’t apply to you in the same way. Your market is more defensible than it looks.

Result: the markets that were ‘too small’ are now worth billions

Horizontal SaaS had its moment. The winners were built in an era when being general-purpose was an advantage, when the path to scale was clear, and when venture capital rewarded growth above all else. That era is ending because the conditions that made it dominant are shifting.

The next wave of valuable software companies will be vertical SaaS businesses that combine deep domain expertise, embedded financial services, and AI-enhanced product development. These companies will have better unit economics, higher retention, and more defensible moats than the horizontal platforms that preceded them. Many of them already exist—they’ve been the underdogs for years, quietly building sustainable businesses in markets that VCs didn’t think were big enough.

If you’re one of those founders, the question isn’t whether the landscape has shifted in your favor. It has. The question is whether you’re positioned to capitalize on it. Are you thinking about fintech as a core part of your business model, or just as a nice-to-have feature? Are you using AI to accelerate your product development in ways that your competitors can’t match? Are you articulating the value of your domain expertise in a way that makes it clear why you’re not just another SaaS company, but the definitive solution for your industry?

The markets that nobody wanted 10—even 3—years ago are now worth billions.

And most importantly, vertical SaaS is the best place to build AI agents solving real-life-problems for the whole SMB economy. I’ll cover that in a second article.

Now is the hour of the underdogs. Make sure you don’t let it pass you by.

To get the Hexa Roundup delivered every month to your inbox, don’t forget to subscribe below - or go to your account settings to turn of Hexa Roundup emails.

Subscribe now

🔮 The Hexa partners’ 29 bets for 2026

Every year, the Hexa partners compile their predictions for the forces shaping tech and entrepreneurship. This year’s main opiniçon: 2026 is when the battles get decided. After two years of AI exuberance, the shakeout begins, separating the lasting from the temporary. The team covers everything from why distribution will beat intelligence, to Europe’s path on climate and sovereignty, to the end of fake news.

→ Read all 29 predictions

💼 Building 40 apps and making 0 apologies, with Jérémy Goillot

Jérémy Goillot is building not just one app, but 40, and he’s very clear about what that costs. In this Start Mode episode hosted by Mat Vaxelaire, the Mobile First Company CEO doesn’t hold back: on firing his cofounder two months in, on why working twice as hard isn’t optional, on the reality that something always has to give. 18 months in, €12M raised, 5,000+ customers, and a move to the U.S. A rare conversation with a founder who’s building at full speed and honest about what it takes.

→ Watch the episode

🎯 How to scale from one-man show to executive committee

Most bootstrapped founders hit the same wall around €2M ARR: they’re doing everything, surrounded by junior people, and suddenly the behavior that created value starts destroying it. In this piece, Augustin breaks down how Veevart, Hexa Scale’s first company, went from founder-led to executive-led - hiring five C-levels in nine months. The hardest parts were getting Antonio (the founder) to believe delegation was possible, learning when to say no to “good enough” candidates, and turning A-players into a team that solves problems without the founder in the room.

→ Read the breakdown

Portfolio news

🩺 DermaScan raises €2.5M to expand early skin cancer screening

A big well done to DermaScan (HX24), who has raised €2.5M to make skin cancer screening faster and more accessible. Cofounded by Florian Legris and Hadrien Lepage within Hexa Start, DermaScan built a dedicated first medical screening layer that enables earlier detection and faster referral to dermatologists. In just one year, the team screened 5,000+ patients, detected 50 cancers, and built a fully integrated medical stack. Next: opening new screening centers across France. Book your appointment.

🛰️ Rubicon joins Hexa Sprint to see what traditional radar can’t

A warm welcome to Rubicon, joining Hexa’s Sprint program in the defense cohort led by Guillaume Lerouge. Founded by Alexandre Boissier and Ines Meraoumia, Rubicon builds AI-powered radar sensing to see beneath forests, soils, and concealed environments in near real time. The team combines low-frequency radar data and AI to detect and classify objects where traditional sensing falls short, giving NATO countries and their allies a decisive edge in image intelligence. Find out more about the cohort.

☀️ Radiant joins Hexa Sprint to replace fossil fuels in industrial heat

Industrial heat alone generates 15% of global CO₂ emissions. Radiant is rethinking how industries produce it, using the power of the sun. Cofounded by Thomas Delhon and Alexandre Meurisse within Hexa Sprint’s Carbon Zero cohort, the team built proprietary solar concentration technology that delivers carbon-free heat from 200 to 1,000°C at a cost competitive with fossil fuels. Find out more about the cohort.

🚀 Swish launches to stop B2B teams losing conversions to timing

Early-stage B2B SaaS teams already have the data to know who’s ready to convert, they just can’t act on it fast enough. Swish, cofounded by Nicolò Magnante and Arseniy Shishaev in Hexa Start, plugs into existing stacks (PostHog, Mixpanel, Snowflake), identifies conversion-ready users through product signals, and automatically sends personalized messages that drive revenue. No engineering time, no manual tracking, no more missed conversions.

🤓 Verso goes live to speed up consumer research with AI

Consumer research is slow and expensive, blocking decisions instead of enabling them. Verso, cofounded by Lydia Bellahouel and Camille Epitalon in Hexa Start, delivers qualitative research in days through AI-moderated video interviews with automated synthesis. The team is seeing strong early traction: high perceived value and recurring usage from month one. The goal is to turn research from a one-off project into a continuous decision engine.

Hexa news

💚 Maxime Brousse joins Hexa Sprint to lead Carbon Zero cohort

Maxime Brousse is joining Hexa as Partner to launch Carbon Zero, our climate tech cohort within Hexa Sprint. Maxime brings 10+ years building Selency (raised ~€50M, scaled to ~70 people) and investing in 70+ startups as angel and Sequoia Scout. Carbon Zero backs climate tech built for systemic replacement that outperforms legacy systems economically, on cost, performance, and scale. €500K for ~10% equity with hands-on support, with one goal: 10x your company’s chance of success. Learn more.

🚀 Hexa Start updates its deal from 30% to 25% equity

Hexa Start is moving to even more founder-friendly terms: 25% post-seed equity instead of 30%, with the same level of support. With AI making it faster and cheaper to validate ideas and ship product, ambitious founders have more options than before. We want Hexa to be the obvious cofounding choice. What you get: a validated ambitious idea, a matched top-tier cofounder, 12 months of costs covered, a squad of experts to build with you, 3 months in San Francisco, and a fast track to seed at above-average valuations. Learn more.

Jobs

• Founding Engineer at Brokerforce

• Talent Growth Engineer at Hexa

Thanks for reading! Follow us on Linkedin and Instagram to stay up to date between issues.

Cheers,

The Hexa Team