The Convergence Nobody's Talking About

A week ago I wrote about the accelerating shift toward Consultancy and agency-owned AI platforms, and why the real story wasn't about services firms becoming software companies, but about who controls the layer where strategic decisions get made.

The response was interesting. Most people agreed with the diagnosis. Several added examples. A few pushed back and that was appreciated.

But there's a dimension of this shift I deliberately left out of that first piece, because I wanted to make the case for it separately. Here it is:

The conversation about agencies building platforms is only half the picture. At exactly the same moment, the major technology platforms are building service teams.

Hyperscalers. MarTech vendors. AI platform companies. They are embedding professional services capabilities, building managed service wrappers, and positioning themselves as end-to-end operational partners. The convergence isn't unidirectional. It's happening from both sides simultaneously and both sides are moving toward the same thing: the decisioning layer, and the client dependency that comes with sitting inside it.

I don't raise this to be alarmist. The capabilities being built are genuinely impressive, and the integration on offer is real. But I think enterprise buyers are currently being asked to evaluate a one-sided version of this dynamic, and that creates risk.

The question most vendor conversations are designed to answer is: does this platform solve my problem? But I think that's the wrong question to be leading with. The question that matters more over a three-to-five year horizon is: what does my operating model look like if I ever need to move away from this?

Very few vendors have a good answer to that question. Most are incentivised not to.

This is where I think the market is heading toward a meaningful crossroad. On one side, firms (whether their origins are in services or software) that are building models designed to deepen dependency as the primary commercial strategy. On the other, a smaller number of operators whose model depends on delivering measurable outcomes rather than architectural lock-in.

The distinction sounds obvious when you state it plainly. It becomes considerably less obvious when you're being shown a compelling platform demo and a favourable commercial structure.

My view is that the firms that strengthen enterprise trust in the next five years won't be the ones who moved fastest to own the stack. They'll be the ones who stayed honest about where their interests ended and the client's began. And who built the depth to deliver on that without using lock-in as a crutch.

That's a harder model to sustain. But it's the right one to be building.