Recently, I moved from Berlin to Munich, so we had a little farewell Stammtisch in a Berlin steakhouse.

(By the way: Stammtisch is not a networking format. We barely talk about work. But sometimes, as it tends to happen in the Berlin tech bubble, we end up talking about how AI is ripping up the organizational playbook.)

One of the guys mentioned that his C-level encourages business unit leads to vibe code their own solutions with AI. They sit closest to the problem and often know what they need better than any backlog process can capture. I had a response ready — I’d been carrying it around for months: AI collapses Product and Engineering into one department. The scarce skill is translating business requirements into specs. Put both in the hands of the people closest to the problem, and you cut the handoff.

But I hadn’t fully understood what happens next.

I’m seeing the same thing across almost every organisation I work with. Most are pushing AI aggressively — they’re tech companies, after all. And yet, many CEOs still seem to think the real impact is somewhere down the road.

It isn’t. It’s already here.

And it goes well beyond Product and Engineering: if business units can build their own tools, what happens to the teams that used to build them? If AI can handle much of the analysis, documentation and coordination around that work, what happens to the layers built around it?

Some of you will roll your eyes and say: obviously. But I’m not sure we’ve really grasped the scale of what’s happening — and that it is happening right now.

The first shift is obvious: Product and Engineering increasingly collapse into the same function as the people closest to the problem can both specify and build solutions themselves.

The more important shift is what happens to Product and Engineering after that. Their work moves from building things to advising and enabling the people who build them.

Product becomes less about owning the roadmap and more about helping the business make better decisions about what to build and why. Engineering becomes less about owning delivery and more about creating the components, standards and guardrails that let others build safely and quickly.

That sounds like a role change. It is much bigger than that.

The nature of the work changes. So does the mindset of the people doing it. Engineers who took pride in shipping features now need to take pride in making someone else capable of shipping. Product people who owned the roadmap now need to create clarity without owning every decision.

That, to me, is the real transformation work: redesigning the software value chain around a world where far more people can build.

It’s a heavy lift.

Pick one piece of software your team built recently. Map the journey from business problem to production.

Then ask, at every step: Why does this need to involve another team?

If AI means the person who has the problem can now specify and build the solution, one of those steps may simply be organisational residue.

Remove one this quarter.

Yours,