Ask almost any engineering leader and they will tell you their people are faster with AI. They are right. The studies hold up, the demos are real, and the adoption dashboards are green. Then the same leader looks at delivery predictability, or at the income statement, and finds that very little has actually moved.

This is the paradox sitting inside most AI programmes right now. McKinsey reports that around 80 percent of firms have adopted AI, and about the same share have seen no meaningful impact on revenue or profit. Mass adoption, real individual gains, and negligible institutional return. Something between the faster individual and the unchanged business is absorbing the difference.

That something is the operating model, and it is the part almost nobody has touched.

The faster individual is already here. The faster company is a design decision. The signature insight · Operating Model

Why faster individuals do not make a faster company

There is a precedent worth knowing. When factories first swapped steam engines for electric motors in the 1890s, productivity barely moved for thirty years. The owners had installed the new power and kept the old floor plan: the same lines, the same sequence, the same workflow built around the steam shaft. The breakthrough came a generation later, when factories were redesigned around what electricity actually made possible. As one analysis of that moment puts it, the factories that electrified first lost to the ones that redesigned the floor. The same analysis draws the blunt conclusion for today: productive individuals do not make productive firms.

Most companies are running the 1890s version with AI. They have bought the motor and kept the factory. The individual is faster, but the work still flows through ownership, decision rights, dependencies and hand-offs designed for the pace before. Only about a fifth of organisations using generative AI have redesigned any of their workflows. The rest have bolted a faster tool onto an operating model that cannot carry the speed, then wondered where the speed went.

What the pattern actually looks like

What to design instead

The lever was never the tool. It is the system the tool runs inside, and it has to be redesigned on purpose.

The takeaway

The AI productivity paradox is not a technology problem, and a better model will not solve it. The faster individual is already here. The faster company arrives only when the operating model between them is redesigned to carry the speed. Swap the motor if you like. It changes little until you rebuild the factory.