The part of consulting that AI has genuinely disrupted is the synthesis layer. Research compresses. The production of a credible diagnosis now takes hours where it once took weeks. What has not compressed is the work of getting an organization to actually change, which the source frames plainly as a contact sport that only humans can play.
Where AI sits in the consulting stack
Recommendation generation is fundamentally a data and reasoning problem. Given sufficient inputs, a model can produce a strategic diagnosis as credible as what a team of analysts would deliver after a full engagement. The deliverable itself is largely solved. Whether anyone acts on it is a different question entirely.
The economics of this split are worth tracking. The research and synthesis phase is where junior-analyst hours have historically accumulated. If that phase compresses, the fee structure of a traditional engagement compresses with it. The value that remains concentrates at the end of the process, in the room where the client actually has to change something.
The piece that does not automate
Implementation is not an information problem. A company receiving a recommendation to restructure a division or overhaul a pricing model often already knows what it should do. The obstacle is organizational. The people whose budgets or reporting lines are affected will push back, and that pushback plays out in meetings and senior relationships that no tool currently navigates on its own.
Staying in the room through that resistance, adjusting the argument, building trust across a leadership team, is the contact sport the source identifies. The person doing it has to be present, reading the room in real time.
What this leaves for the industry
The firms that use AI to compress the research phase while protecting the implementation relationship are in a structurally different position than those treating it as a replacement for delivery itself. Compressing low-value hours is efficiency; replacing the relationship is a margin gamble.
The constraint is organizational physics. People change when other people are in the room persuading them to. That is the one piece of the consulting stack that has not moved.