I Built a Company on the Side. My Chief Growth Officer is an AI.

I want to be clear about something upfront: this is not my main business. I run a consultancy. I work with tech companies, e-commerce brands, gaming platforms, helping them grow, optimize their CRM, get more out of their data. That's my day job.

This is what I built on the side.

I had the idea for a while. An AI-powered platform, something I believed in, something I kept coming back to. But like most ideas, it was sitting there waiting, for time, for the right moment, for something to finally push me over the edge.

That push came from an unlikely place: I just decided to ask Claude.

Not a co-founder. Not a consultant. An AI. I gave it my rough ideas, my instincts, my angle, and then I said: if you had free rein to build this, how would you do it?

What came back floored me.

In the span of three hours, I went from a concept to a full business plan. Not a summary. Not a list of suggestions. A real plan, with brand strategy, character design, creative direction, visual asset prompts for generative AI tools, technical specifications, and a complete product roadmap. Three hours. I've been in this industry for a long time, and I've never seen anything move that fast.

The platform itself, I'm not ready to name it yet. But trust me, it's related to AI. And yes, the irony is not lost on me.

Then we started building.

I gave Claude Code access to everything, front end, back end, the full stack. It wrote the code, generated all the copy, designed the prompts for every visual asset the platform needed, called external APIs to produce those assets, and uploaded them directly to our system. The loop was fully automatic. I wasn't just working faster. I was working in a completely different way.

The prototype was live within days. No team. No office. No standups. Just me, a few hours a week, and an AI that doesn't sleep.

But here's where it gets really interesting.

I had built the platform to be data-driven from the start, event tracking, behavioral data, conversion checkpoints baked in from day one. And I had a decision to make about what to do with all that data.

I connected Amplitude. Took about ten minutes. And then I gave Claude something I don't think many people have given an AI before: a proper title.

Chief Growth Officer.

This wasn't symbolic. I gave it full access to our analytics, user behavior, drop-off points, engagement patterns, the entire funnel. And I told it to act like it owned the growth function. Identify blockers. Generate hypotheses. Design experiments with proper control groups. Prioritize a backlog. Close the loop from insight to deployment as fast as possible.

What it did was exactly what I do professionally for my consulting clients, the full cycle from measurement to hypothesis to specification to code to deployment. Except now I wasn't doing it manually. It was running. On a side project. In the hours between my actual work.

The results came fast.

Our platform's core experience is built around chat, interactive, AI-driven conversations that drive the whole engagement loop. In the early days, our conversion rate from chat to lead was sitting at around 5%. Decent, but not where I wanted it.

Claude analyzed the data. Spotted the friction. Proposed changes. I reviewed them, asked the right questions, applied them, always with a human in the loop, always with my own judgment in the mix. But the velocity was unlike anything I'd experienced.

Within a single day, conversion climbed to between 20 and 25 percent.

I've been working in growth for years. I've never seen a jump like that, that fast.

When people ask me whether it was threatening to watch an AI do what I've spent years doing professionally, my answer is simple: it was beyond exciting.

What I felt, and still feel, is that we are in the very early stages of shaping an entirely new methodology for growth and product development. The first wave of modern growth was about unifying analytics, connecting market data, acquisition data, and product behavior into one coherent picture. That was already powerful. But what's happening now is something bigger: we are finally closing the gap between insights, marketing, and actual development. All three are now connected in a single, continuous loop.

The old paradigm was the Lean Startup, learn fast, fail fast. The new paradigm is learn automatically and instantly. The iteration cycle has compressed from weeks to hours, sometimes minutes.

I don't fully know where this is taking us. But my instinct is that we are heading toward hyper-complex products, things that were simply not realistically possible to build before, at least not without large teams and long timelines. And that is both thrilling and a little frightening. Because these super-complex products become harder to fully understand, harder to debug, almost like they develop a logic of their own that goes beyond what any one person can hold in their head. New frameworks for how to work, how to govern, how to make decisions, they will need to be invented. The classic rituals of standups and two-week sprints? That feels like 2005. We need something new.

People ask me how I know when to trust Claude and when to push back. Honestly, I don't obsess over whether any individual idea is right or wrong. Everybody has bad ideas sometimes, human or AI. What I care about is whether the method is sound. Is the experiment designed correctly? Is there a proper control group? Is the A/B test statistically valid? Are we reading signal or just noise? If the scientific method is solid, I trust the process. The outcome takes care of itself.

As for my own role, I'm not sure there's a clean job title for what I do now. The closest analogy I have is a board member, or a senior advisor. A sounding board. Someone who nudges the AI in directions when intuition and experience say something matters that the data hasn't caught yet. Because here's what I know after doing this: human beings are still remarkably creative. We still connect dots in ways that AI doesn't always anticipate. That edge, that instinct, is where I show up.

My working day on this project is a continuous flow. A real-time experiment that never fully stops. Ideas moving back and forth, data being pulled and interrogated in different ways, some hypotheses tested immediately, others left to simmer. It's less like a side project and more like a living system that I'm part of, steering, questioning, occasionally overruling, but mostly just trying to keep up.

Now, I understand not everyone is going to see this clearly. If you are deeply attached to your rituals, your two-week sprints, your backlog grooming sessions, your carefully choreographed standups, then yes, you can keep doing that. I've spent years working with tech companies and I understand exactly why those frameworks exist and what problems they were designed to solve. I respect them. I've used them.

But we have crossed a threshold. We are living through a quantum leap, a moment where the rules of how you build, how you grow, and how you compete have fundamentally changed. Not incrementally. Fundamentally. And not everybody is going to see it right away, and that's okay. But the ones who do are going to move at a speed that makes everything else look like it's standing still.

If I could go back to the moment before I started all of this and tell myself one thing, change one decision, avoid one mistake, do something differently?

I wouldn't change a thing.

I'm mind blown.

The Chief Growth Officer has not yet asked for equity. I'm keeping an eye on that.