Overlap

I was already working this way.

The overlap has always been the point. AI has made it much cheaper to move between strategy, design, and build.

01: Judgement

The handoffs got cheaper. Judgment matters more.

For most of the past decade, product teams divided work neatly: PMs wrote specs, designers made them usable, and engineers built them. Every handoff cost time and context, but the model worked well enough to become standard.

AI has lowered the cost of moving between those roles. It has not replaced them. Someone who can think across strategy, design, and engineering can now take an idea much further before it needs a handoff.

You can see the change in ordinary work. Design review happens in a PR thread. Agents turn briefs into working prototypes. Product judgment matters in the brief, the eval, the prototype, and the decision to stop or keep going. The advantage is no longer just output. It is knowing what deserves to be built.

This is not a theory. It is your Wednesday.

“Who writes the brief? That is the question that matters. Not who builds the thing, but who has the judgment to direct it.”

That judgment is not new. It was often buried under coordination work. With cheaper handoffs, it shows up in what ships, what gets cut, and whether the product decision still makes sense once it reaches code.

02: The new scarcity

A director holds the whole film in one mind.

The film director analogy explains what gets scarce when moving between roles becomes easier.

A director is the only person on a production who holds the entire film simultaneously. The cinematographer is an expert in light. The composer is an expert in score. The editor is an expert in rhythm. Each is good at a particular discipline. The director is not necessarily better at any one of them, but is the person responsible for how they work together. If the lighting is technically right but wrong for the scene, the director catches it. That view of the whole is the job.

When strategy, design, and engineering sit in the same head, decisions can happen together instead of one after another. Less gets lost in translation. The product is judged as one system, not passed from room to room. AI makes that range more useful if you already know how the parts fit together.

“When something isn't working, you don't need three separate conversations to find it. You can see where the system is wrong.”

A director creates the conditions for good work. The designer gets a clear intention rather than a guess. The engineer gets decisions that hold. Problems show up at the seams before they become expensive. Each specialist can spend more time on work that needs their expertise.

There is another part of the job. A leader who works across strategy, design, and engineering, and uses AI in production, can also redesign how the team works. They decide which workflows belong to agents, where people need to stay involved, and how to build evals that catch regressions. You cannot outsource that redesign to someone who has not worked inside the new model.

I have rebuilt a team's operating model before. During eBuddy's partnership with Gree in Japan, I changed a top-down development process into an agile cadence across teams in Amsterdam and Tokyo. Velocity rose 10×, while quality and predictability improved too. The instinct is the same now: notice when the system no longer fits the work, then change it.

The more of the product you can see, the more useful AI becomes.

03: The record

I've only ever built on the whole product

I have worked on voice UI before the category had a name, built marketplaces from scratch, and helped products reach millions of users. There have been two exits. The through-line is strategy, design, and build held close to the same product. AI did not create that way of working. It gave it more reach.

In 1999, I co-authored The Piranha Economy, a Dutch bestseller (Addison Wesley) about how internet technology was reshaping business. Then I went and tried to build it. I co-founded XOIP, Europe's first unified messaging company. It was named the Netherlands' #1 hottest startup and eventually sold to a European ISP. In 2003, inside that company, I prototyped a voice-driven GUI in a markup language almost no one remembers. There was no market for it and no job title for the person building it. I built it because I could see where it might go. That has always been reason enough for me to build something.

Next came Cerbex, a hotel-group-booking auction marketplace for Chinese tour operators. It was a two-sided marketplace with a new pricing model in an industry I had not grown up in. Annual revenue grew from €10k to €9M.

Then eBuddy. I led the pivot from a popular IM client into a multi-platform messaging app: 0 to 7M users, #1 in App Stores in several countries, three engineering teams plus design. The piece I'm proudest of is the process rebuild that came with the Gree partnership in Japan. We shipped 10× faster without losing quality or predictability. I co-founded two startups and saw two exits. Neither followed a map.

“When the early ChatGPT API dropped, I didn't read about it. I started building. I understand new technology by using it before the tooling arrives.”

That instinct is clearest in the XOIP work today. When AI began making our existing call-routing product obsolete, patching it was the wrong answer. An AI-native interaction needed a product designed from scratch. I made the strategic call, designed the experience, and wrote production code. The featured case study shows the result.

04: If this resonates

I'd like to hear about what you're building.

If you are building or rebuilding a product team for what comes next, and need someone who can lead across strategy, design, and what ships, I'd like to talk.

I am most useful as a Head of Product or first senior product hire on a team reshaping itself around AI. I bring clearer direction and experience working in the model you are trying to build.