So the problem is not just that an AI model cannot do the face-to-face work of building consensus behind a plan. It’s that the dataset that would be needed to train an AI to design the plan does not exist; it only comes into existence when humans engage in the planning struggle. Garicano compares those who think AI can solve this sort of problem to the socialist central planners who – as the economist Friedrich Hayek pointed out – wrongly believed they could do without the local knowledge revealed in market prices.
To sum up, in the age of AI, strategic consultants don’t actually sell intelligence at all. They sell a process.
Even if you believe this argument, the strategic consulting industry seems certain to get smaller. Analytic and technical work was the bulk of the meal the industry once served to customers, while the bespoke process of helping management adapt to change was the special sauce. Now there is just sauce.
The challenges will be how to train young consultants when there is not much bulk analytic work to be done, and reforming the fee model. On the latter front, change has begun. More firms are charging not for their time, but by desired outcomes. This is the future. Per diem work is on a burning platform.
Doctors are notoriously bad at taking their own medical advice. It will be interesting to see how the strategics do at transforming themselves. Consultant, advise thyself.
