Christophe Hébert: “What if we're only at the beginning?”
He took apart his parents' first computer, sold his first IT services at 19, realigned his entire career path in two days, told a client they had budgeted 500,000 euros too much, learned PHP by asking what the arrow was for, then built a recruitment firm and a software product. Over more than two hours, Christophe Hébert told me his story, from his earliest obsessions with computers through to Bluecoders, Marvin, and the turn he is taking today with AI.
Christophe has always loved computers.
Not just writing code. Understanding how things work, testing, taking them apart, starting again. And soon enough, above all, using technology to solve other people's problems.
Long before Bluecoders. Long before Marvin. Long before he even knew he would make a career of it.
As he tells me his story, I keep noticing something that returns at every stage: an almost obsessive curiosity about technology, the urge to understand the system he is operating in, and then very quickly the urge to work out how to do it better.
And the story starts well before he founded his first company.
Understanding everything
The first computer arrives at his parents' home when Christophe is still a child.
Very quickly, he wants to try everything.
When he opens a menu and finds ten options, he clicks all ten just to understand what they do.
Inevitably, sometimes the computer stops working.
At home, whenever there is a problem, the reaction becomes almost automatic: what has Christophe done now?
But that is how he learns.
He discovers software, Photoshop, networks, servers, video encoding, then programming. Not because he methodically decided to learn computing, but because he constantly wants to do something his computer cannot do yet.
One day, his mother asks him again how to find her photos.
So he builds her a small application.
A very simple home screen, with shortcuts that let her go straight to her photos or to the software she needs.
"That's how I discovered programming."
At the time, he does not even feel he is learning something that could have value.
And yet, very young, he is already selling small IT services: setting up computers, putting machines on a network, building systems that let people share files.
At 19 or 20, he works with several real estate agencies and a medical practice.
He also gives tutoring lessons.
Already with something that will resurface much later in his career: the pleasure of understanding something complex, then making it far simpler for someone else.
Two days to realign his career path
At that point, though, computing is not yet an obvious professional path.
Christophe even considers becoming an airline pilot for a while, like his grandfather. In engineering school he moves towards industrial systems management and plans to go to Madrid to study mechanics during his Erasmus year.
Until an exercise where he has to present his career plan.
He does not really have one.
The facilitator in front of him pushes him to think about what he actually enjoys doing.
And the answer suddenly seems fairly obvious: computing.
Why is he organising his whole path around something else when this is the subject that has fascinated him for years?
That same evening, on his way to a party, he talks to a student who tells him he had dreamed of going to Madrid to study mechanics.
The problem: he has been given a course in the Netherlands on computing applied to business.
Christophe has just been given Madrid.
They swap places.
The next day, they have the change approved by their school.
Christophe also switches his major to computing.
"That's how I realigned my career plan in two days."
And this time, something changes.
In class, computing feels far easier to him than to many other students.
But above all, he understands why.
He has already been practising for years.
He had simply never told himself that this ease could be a professional skill.
Realising he isn't just "the geek"
It is during his internships that Christophe really starts to grasp what he can do.
At a web agency, he is handed a project on automated testing.
He moves much faster than expected and, when the project manager goes on holiday, the team offers him the whole subject.
He is still an intern, but he quickly finds himself in client conversations, notably with Veolia, and in direct contact with the agency's leadership.
What he mostly takes away from that period is the trust he is given.
Nobody tells him he is too junior.
They tell him instead: you know how to do this, come along.
And over time, he starts to believe it himself.
His next internship gives him something else.
Christophe joins an organisational consulting firm.
The idea of the job appeals to him a lot: understanding how a company works, improving its processes, increasing its productivity, reorganising things and building the IT tools that go with it.
The reality disappoints him more.
He discovers, among other things, a way of working where a task estimated and sold as two days can sometimes be finished in an hour… without that changing much in the schedule, since the two days have already been sold to the client.
But he also meets a supervisor there who will profoundly change the way he sees himself.
Until then, Christophe has mostly identified with his ease around computers.
A geek, in a way.
His supervisor points out that he also has a real ease in building relationships with people.
"I had never really considered myself good with people."
It matters in his story because he starts to understand that he may not have to choose between the technical and the human.
He can understand a problem, explain it, persuade, bring several stakeholders together around a decision.
And that combination is about to become very useful.
Telling the client he doesn't need his whole budget
It is at MAIF that Christophe really starts to measure what that perspective can bring.
He arrives on a project that is struggling to move forward, with many stakeholders to align and a budget he puts at around a million euros.
He starts by doing what he is asked: understanding the problem and getting the various people back around the table.
Then he looks at what actually remains to be done and the time available to do it.
And reaches a conclusion that is fairly counter-intuitive for someone working in consulting: there is far too much money allocated.
About 500,000 euros too much.
That is anything but trivial.
The firm's model rests precisely in part on the time sold to the client. Christophe has just discovered for himself that work billed as two days remains work billed as two days, even when it is finished much faster.
At MAIF, he takes exactly the opposite reasoning.
If the project does not need all that money, why try to consume it?
So he explains that, in his view, half the budget cannot be used sensibly in the time remaining anyway.
And that is precisely what changes the way the client sees him.
Christophe remembers that they probably were not used to a consultant turning up to tell them they had budgeted too much.
MAIF then asks him to go further.
His role shifts towards project portfolio management: looking at the various budgets, identifying oversized envelopes and reallocating the money to the projects that genuinely need it.
And so he finds himself, very young and still an external contractor, taking part in discussions at a level of responsibility he never imagined reaching so quickly.
That experience also teaches him something else: how the consulting market actually works.
How a company buys time and skills.
How a consultant is valued by a client.
How much their work is billed for.
And inevitably, Christophe also starts to look differently at his own place in the system.
At the end of his internship, a manager on the MAIF side asks whether he intends to accept the permanent contract his firm is due to offer him.
Christophe does not know the offer yet.
So his contact makes him his own: keep working with them at around 600 euros a day, roughly 200 days a year, with a suitable contracting arrangement.
A few days later, his firm offers him 28,000 euros a year on a permanent contract.
Christophe immediately connects the two figures.
On one side, a client valuing his work at around 120,000 euros a year.
On the other, an offer of 28,000 euros in salary.
He tries to discuss it with his firm.
Not because he has suddenly discovered that a company has to make a margin on the work it sells.
But because the gap seems hard to understand given the value the client itself has just put on his work.
And he hits a wall.
"I thought: what kind of world is this, where everyone wants you for how effective you are, but nobody wants to pay you?"
He will not choose to stay at MAIF, though.
The potential pay is very attractive.
But the job does not appeal to him enough.
What interests him more and more is software.
And during that same period, he is already sketching one out.
His idea: letting a company assemble a project team based on the skills available.
A kind of FIFA applied to teams: pick your IT expert, your project manager, the different skills you need, and build the right team.
At this stage, it is mostly a mock-up.
But the problem that will interest him later with Marvin is already there: understanding available skills better and helping companies put the right people in the right place.
And above all, Christophe starts to understand the role he wants to hold.
"Me, I want to be the one who designs the software. That's what interests me."
"What's the arrow for?"
So he starts looking for roles that bring him closer to software publishing.
Several leads come to nothing.
So in the end he calls back the web agency where he did his first internship.
He is looking for a technical project manager role.
They do not have one.
They do, however, offer to make him a developer.
They work in PHP and JavaScript.
Christophe knows neither.
He tells them.
The answer is simple: never mind, he starts on Monday.
When he arrives, he asks a question that will become a running joke in the team.
In PHP, he sees an arrow symbol everywhere.
He asks what it is for.
That is where he starts.
Then he works.
Days with the team, evenings on the train, sometimes at night.
He buys books, learns the basics and gradually wants to understand the whole system.
A few months later, he is working on cinema platforms with significant traffic and availability constraints.
But even as a developer, he does not only look at the code.
He starts wondering why the work is split up the way it is.
How to organise tasks better.
How to reassure a client when a project is not moving as planned.
How to explain clearly what is happening technically.
He keeps coming back to that position between technology, organisation and people.
And in parallel, he starts selling again.
A freelancer mentions a possible website to him and suggests a price of around 2,000 euros.
Christophe meets the client.
Assesses the need.
And judges that there is far more work and value involved than expected.
He ends up selling the project for 12,000 euros, design included.
From then on, a calculation starts running in his head.
"I make more money at the weekend than during the week"
Christophe likes the company he works for.
He likes the people.
He is learning a huge amount.
But he is finding the employment framework harder and harder to live with.
Not only because he wants to be free.
Mostly because, in parallel, he is discovering what he can generate on his own.
And the ratio between the time he gives his job and what he earns starts to frustrate him.
So he goes to his boss with a proposal: move to three days a week.
The way he explains it is fairly direct:
"I make more money at the weekend than during the week. So I need to change the ratio of week to weekend."
His boss says no, partly for organisational reasons.
But he also understands that Christophe wants something else.
His salesperson happens to be leaving.
He offers Christophe the role.
Christophe accepts.
And his first instinct as a salesperson is, in the end… to do computing.
The CRM strikes him as bad.
He installs another one, migrates the data, automates some processing and above all starts making use of years of sales history that nobody really uses any more.
Where others would start from scratch, Christophe looks for the old contacts.
The projects discussed.
The companies already approached.
The needs that might come back around.
He uses technology to choose better where to put his sales energy.
Then he calls.
A lot.
On some days, he reckons he spends three or four hours on the phone.
The meetings start stacking up.
And he discovers that he genuinely enjoys selling.
He learns how large companies buy IT, how projects come about, who decides, where the budgets sit.
But the more he sells, the more a new problem appears.
The company no longer has enough engineers to deliver the projects it signs.
So his boss asks whether he could try to recruit some.
Christophe opens LinkedIn.
Contacts a few developers.
Tells them about the projects.
And after two days, his reaction is very simple: he thinks it is brilliant.
Recruitment: finally a job that brings it all together
A few days later, on a plane, the woman next to him mentions her partner, a headhunter.
Christophe starts asking questions.
He discovers a job where you spend your day searching, calling, understanding a market, persuading people — with pay that can be strongly tied to performance.
It looks a lot like everything he has just discovered he enjoys doing.
He gets back from his weekend.
Redoes his CV.
Applies to Mobiskill, a firm specialising in tech profiles.
A few days later, he is hired.
And for a while, he really does feel he has found the job that fits him.
Technology.
Sales.
People.
Competition.
And a much more direct link between his performance and his results.
The market is buoyant and he also inherits a patch where a lot of groundwork had already been done before him.
He starts fast.
One placement in the first month.
Three in the second.
Then five in another month.
He estimates he generated around 250,000 euros in revenue over five months.
Above all, he loves the job.
But once again, even at the height of his sales performance, he comes back to the tools.
He finds recruitment software bad.
To find a candidate again, searching the notes app on his Mac sometimes feels more effective than the database supposedly built for exactly that.
So he starts building his own tool.
This time, the idea is much closer to Marvin.
Software where the recruiter genuinely works.
That helps them find information again, make use of the candidates they already know, track their activity and understand where to focus their attention.
He presents the project to Mobiskill's leadership.
They find the idea interesting and talk about building something together.
Christophe waits.
Follows up.
But nothing materialises.
Meanwhile, his results keep climbing.
Until the moment when several commissions are due in the same period.
His boss asks for his pay to be put on hold before speaking to him.
Christophe first assumes he is going to be congratulated.
The conversation goes another way.
Part of the calculation of his pay is called into question.
The disagreement ultimately concerns a fairly small amount relative to the revenue he has just generated.
But for Christophe, the problem lies elsewhere.
He thought he had finally found a model where his success depended directly on what he managed to produce.
And he discovers that the moment that success becomes significant, someone can still step in and change the rules.
"My success has to be dictated by the market. Not by some guy putting a lock on me."
That evening, he goes home.
And understands that he no longer wants to depend on that situation.
The next day, he resigns.
This time, entrepreneurship is no longer just a possibility.
The moment when there was still nothing
After Mobiskill, Christophe negotiates a settlement agreement and finds himself with 22 months of unemployment benefit ahead of him.
He knows he wants to build a business.
But not yet exactly what.
He is torn between two directions that have been with him for several years already: building software, or doing sales, probably through a recruitment firm.
He also considers other projects.
And above all, he has one fairly specific fear: becoming the person who tells everyone he is about to launch something and who, months later, still has not done anything.
For him, talking about what he wants to build has rather the opposite effect.
The more he talks about it, the more he sees the gap between the ambition and what actually exists.
And the more that gap pushes him to act.
Over the summer, while the project is still vague, he buys a domain name.
topcoder.io.
32.90 euros on his personal card.
Then early September arrives.
His partner leaves for work early in the morning.
Christophe gets up with her and sits down at his computer.
Except he has no client to call yet, no team to meet, no company to run.
He does not even know exactly what he is going to do with his day.
But he still remembers that moment very clearly.
"I didn't know what I was going to do, but I was full of energy."
What strikes him most is the freedom.
He has almost no commitments.
No employees.
No company whose existing business he has to protect.
Very few personal costs.
"The only thing I had in my life at that point was my rent. So if I paid my rent, I was living."
And it is precisely a moment he finds rarely told in entrepreneurial stories.
"In entrepreneurship we talk a lot about all sorts of moments. We don't talk about the moment when you don't know what to do yet and you're already sitting down at a table."
When I ask him what he actually imagined that day, his answer is ultimately fairly simple.
He was torn between doing software and doing sales.
Years later, when he looks at Bluecoders and Marvin, that is precisely what makes him proud:
"Today, we do both."
That does not mean he considers he has achieved what he imagined that day.
Bluecoders has not reached the level he wants to take it to.
Nor has Marvin.
But the two directions he could see ahead of him that morning now genuinely exist.
Bluecoders: experts recruiting experts
At the start, the vision for Bluecoders is closely tied to his own path.
Christophe wants to build a firm where the tools are good enough to make recruiters highly autonomous.
And because his engineering training helped him enormously in understanding the developers he recruits, he sets out with one conviction: recruiting experts with experts.
So at first he imagines hiring mainly engineers and turning them into recruiters.
Reality gradually makes him adjust that idea.
Understanding a technical job perfectly does not mean enjoying calling, selling, persuading or spending your day with candidates and clients.
Some engineers naturally want to go back to the technical side.
But the underlying idea holds.
To be credible in a niche market, Christophe is convinced you have to genuinely understand the people you are recruiting for.
Bluecoders grows fast.
Christophe explains that after roughly two years, the firm was already at around 200,000 euros of monthly revenue.
That level does not feel entirely unfamiliar to him: he had seen a recruitment organisation operating at that scale at Mobiskill, so he knows what it should look like.
The hard part comes afterwards.
Going beyond the level you already know.
As he remembers it, Bluecoders' best month will come later, at around 300,000 to 320,000 euros of revenue.
But the firm does not manage to hold that level durably.
And that is where Christophe discovers the model's real limits.
Growing a business that depends on people
Christophe loves recruitment deeply.
But it is also a business that frustrates him.
Because everything rests on people.
To do more, you need people who can sell, recruit, build a network, maintain their relationships and accumulate experience.
And when those people leave, part of that value leaves with them.
"You're always caught by your own turnover."
Bluecoders can have an excellent month, then start the next one with an enormous amount of work to rebuild.
The market itself adds volatility.
When companies are hiring heavily, the firm can accelerate extremely fast.
When they slow down, the situation changes just as fast.
Over the years, Christophe is therefore less interested in whether Bluecoders can post a very big month than in understanding how to build a model that can keep growing without constantly starting over.
That also explains how the firm has evolved: specialisation, working with freelance recruiters, and today the idea of developing several specialised markets around the same model.
Not to become a generalist.
Quite the opposite.
To be specialised enough in each market to be credible, while being less dependent on a single sector.
He pictures, in time, a group able to operate across several verticals and absorb market cycles better.
His view of success, though, remains very demanding.
When I ask him how far he considers he has taken Bluecoders and Marvin, his first answer is almost the opposite of a satisfied assessment: nowhere near the level he still wants to take them to.
"Scattering yourself is a form of fear"
It is precisely in trying to build several things that Christophe believes he made one of the most costly mistakes of his career.
Scattering himself.
Bluecoders.
Marvin.
Then real estate opportunities, notably around offices, which can be financially interesting but interest him far less.
And other subjects he felt like seizing along the way.
For a long time, seeing an opportunity makes him want to go for it.
Today, his relationship with that has changed.
"At some point in your life, you discover that not every opportunity is worth taking."
With hindsight, if he had to start again, he thinks he would focus on software much earlier.
He would probably also look to find a partner sooner.
And above all, he would persevere more on fewer things.
"I think scattering yourself is a form of fear."
Because doing several things can give the impression of making enormous progress.
But it also avoids putting all your energy and all your risk on a single bet.
He does, however, now draw a real distinction between scattering and strategic diversification.
Developing several unrelated activities scatters you.
Building several verticals around the same business can, on the contrary, strengthen the whole.
He says he now takes more pleasure in turning some opportunities down.
Because he places more value than before on two things: commitment and perseverance.
And that idea goes beyond his own projects.
It also shapes how he looks at the people he works with.
In a business like consulting or recruitment, where turnover can break momentum built over years, he increasingly values people who commit long enough to live through several phases of a project.
The moment AI changes Marvin's trajectory
Generative AI enters this story when Marvin has already existed for several years.
At first, Christophe is curious.
But not shaken.
He tests a lot.
He sees the value in rewriting an email, reformatting information or automating certain tasks.
But he still finds the tools too dependent on context you have to keep feeding back to them.
What interests him most is a new capability.
Until then, IT automation works particularly well when you know in advance the format of what goes in and what has to come out.
AI makes it possible to start automating what varies.
"For me, AI is the automation building block I was missing in IT automation."
So he keeps testing.
Even when he considers it is not good enough yet.
Then something shifts.
For the concrete application to Marvin, he places that change around the start of 2026.
In January, he starts to see the potential.
In February, the team tests more.
In March, they actually apply it to their own development problems.
And there, Christophe reaches a far more radical conviction.
The way software is built is changing.
The problem is that Marvin carries several years of technical history.
Development on the software started around 2017.
Features were added, then removed.
Decisions were made, then changed.
Different people wrote different parts of the product.
Like many older pieces of software, it accumulates technical debt.
And that debt becomes even more of a problem when you want AI to work on it.
AI is powerful.
But it needs to understand precisely the system it is operating in.
The more old paths, inconsistencies and dead parts a product contains, the harder it is to ask AI to evolve the whole thing cleanly.
So Christophe takes a heavy decision: rebuild Marvin.
The team spends several months migrating to a new version.
With a great deal of human effort.
But for Christophe, what is at stake goes far beyond adding AI features.
His definition of an AI-first product is different:
"AI-first software is software you can iterate on with AI."
So the competitive advantage does not lie only in what the user sees today.
It lies in how fast the product will be able to change tomorrow.
When the problem becomes not doing too much
That new speed nevertheless creates a risk Christophe had not experienced in the same way before.
When building something is expensive and slow, the technical constraint naturally forces you to choose.
When delivery becomes much easier, you can start saying yes to everything.
One more feature.
An automation.
A new tool.
A client request.
A new idea.
And end up with a product that has become needlessly complex.
So for Christophe, the problem moves.
The question is no longer only: can we build this?
But: should we really build it?
"The hardest thing in AI today is keeping something clean, staying simple."
He compares AI to an extremely powerful horse that needs blinkers.
The faster it can go, the more precise you have to be about the direction.
And that is exactly where he sees the value of software in a world where everyone will have access to the same AI models.
The software provides the context.
The rules.
The constraints.
AI does not need to rediscover, with every request, who the user is, what they are trying to do and how their job works.
All of that is already around it.
"You can't use AI a little"
But Christophe's main concern today is not only technical.
It is about people.
He does not think the right answer is to replace teams abruptly.
At Marvin, his choice has instead been to train the people already there.
But he believes the gap between those who genuinely learn to work with AI and those who use it occasionally risks becoming very large.
"With this, you can't use AI a little. You have to become an AI expert."
His comparison with Excel captures his reasoning well.
When someone learned Excel two months after everyone else, they could catch up on two months.
Excel had not progressed radically in the meantime.
With AI, during the two months someone spends learning, the tools themselves keep evolving.
So the gap is not only about a skill.
It is also about how fast you are able to keep training yourself.
And for Christophe, that is probably one of the biggest changes coming to many professions.
Exposing yourself to opportunity
That way of approaching AI connects to another idea running through his whole story.
Christophe gives luck far more room in entrepreneurship than you usually hear in discussions about execution.
Not entirely passive luck.
Luck you choose to expose yourself to.
"The hardest part is spotting the opportunity, putting yourself in the right place at the right time to catch the wave."
Execution remains essential.
But it becomes much easier when you are positioned in the right market at the right time.
For him, then, building a business also means putting yourself in situations where something can happen.
Testing a technology before you are sure it will work.
Meeting people.
Building a first product.
Making a decision early enough that the next step becomes possible.
That is also why he thinks a lack of decision can be very costly.
"I think it's better to get it wrong than to do nothing."
Because a decision does not only let you seize the opportunity in front of you.
It can open access to the one arriving right after.
But that is also where his two lessons meet.
Exposing yourself to opportunities does not mean taking all of them.
Today, Christophe increasingly looks to choose the ones that strengthen what he is already building.
Learning to stop carrying it all alone
Christophe launched his projects alone.
The way he operates was long fairly simple: he sees something, starts moving, then finds the people he needs along the way.
He does not wait for someone's validation to begin.
But over the years, his view of partnership has changed.
He has understood that it is not simply about finding people who can do a lot of things.
But about finding people who are particularly strong on what genuinely creates value in the business.
For Bluecoders, that value is deeply human.
Creating a collective.
Making people want to stay.
Making sure the energy does not rest on a single moment or a single person.
That, in fact, is how he talks about our partnership.
He tells me that at Bluecoders, what you really have to be able to do is make people want to be there — and that this is something he considers me better at than him over the long run.
At Marvin, he talks about Benjamin with the same logic.
Not for the same skills.
But because he sees in him the ability to make a complex technology pleasant and simple to use.
He describes him as someone able to make you feel, inside a piece of software, "like you're in a five-star hotel".
With hindsight, then, Christophe considers that some of his best decisions were finding, for each project, someone strong on what genuinely sits at the core of its value.
It is no longer about knowing how to do everything himself.
But about knowing who to build with.
"What if we're only at the beginning?"
At the end of our conversation, I ask Christophe whether there is anything he did not have time to say.
An idea he would like to finish on.
"I think you always have to tell yourself you're only at the beginning."
He talks about the risk, in entrepreneurship, of gradually shifting from creative energy to purely managerial energy.
Of stopping imagining what could still be built, and simply maintaining what already exists.
And that is precisely what he does not want.
"What if we're only at the beginning? What would the end even be?"
What is quite striking after more than two hours of conversation is that he almost never talks about Bluecoders or Marvin as finished successes.
When I ask whether he has now achieved what he imagined that morning he sat down at his computer without knowing exactly what he was going to do, his answer is no.
He is proud to have ended up building the two directions he could already see back then: sales with Bluecoders, software with Marvin.
But for him, both are still under construction.
And in the end, that may be the best way to understand his relationship with entrepreneurship.
Keep building long enough to succeed.
But never consider the creative part finished.
Always being able to ask yourself the same question again: what if we're only at the beginning?