Two PMI models. One identical problem.

There are two opposing answers to the same question: what happens after closing?

Model 1 — Integration. One platform.

One stack. One product. One cost base. The alpha comes from the integration itself. Synergies that most buyers put in the model and never realise, because realising them is hard. Whoever can do it buys a collection and sells a company. The excess return sits exactly in that difference. The price: time, capital, and an organisation that spends two years busy with itself.

Model 2 — No integration. One data layer.

The acquired company's technology stack stays as it is. No migration. No rip-and-replace. Instead, a new shared data layer across all systems. From then on, nobody consults the system. They consult the AI. The alpha comes from pace. You can buy the next company before the last one is digested.

The price: you end up with data sovereignty — but no shared machine. The collection never turns into a product on its own. The two models are opposites in almost everything. Capital deployed. Speed. Risk profile. Exit story.

The one point where they are identical

Both stand and fall with the same people. Because in parallel, the same thing happens in both worlds: a development freeze.

Developers are no longer allowed to build anything on their own. Only with AI. And the reaction is exactly what anyone could have predicted — and still nobody has a plan for it: the younger developers love it. The older ones refuse.

This is not a technology problem. It is a status problem. Someone who was the one who could do it for twenty years does not lose a tool. He loses his place in the pecking order. No training programme in the world fixes that. Whether you merge the systems or leave them standing changes nothing. It only shifts the moment you notice.

This is exactly where a need is emerging that the market has not priced yet. Not the consultant who hands over an AI concept. Not the coach who runs acceptance workshops. But the manager or interim CEO who is responsible for three things at once:

1. Building with AI without layoffs. Whoever lays people off after introducing AI admits he had no plan for the capacity it freed up. Freed developer time is not a cost item. It is a product lead — provided someone redirects it.

2. Change management across generations. Not "taking the staff along". Giving the older ones a role that is worth more than the old one. Otherwise refusal wins. Quietly. Through deadlines that are not met.

3. Building new structures without tearing down the old ones. Leaving legacy systems in place or merging them is an architecture decision. The integration work on the people is needed either way — except that in the second model nobody does it, because everyone believes it disappeared with the stack.

Anyone who thinks they have to choose between model 1 and model 2 is having the wrong debate. Both models work. Both fail at the same point.

If your company is in exactly this transition: write to me.