Sort clients by AI posture and serve both groups deliberately
Client bases are splitting along AI-forward, moving-slowly and AI-averse lines; firms that run a single operating mode for everyone will produce the wrong shape of work for a growing share of their book, and need to classify and serve the segments deliberately.
Until recently, a mid-tier firm could assume a reasonably uniform client base when it came to how work would be consumed. A report was a report. A deck was a deck. The form the firm produced was the form the client expected, and variation was mostly about subject-matter and seniority rather than about AI posture.
That uniformity is ending. Client behaviour is splitting along at least three visible lines. A minority — the AI-forward — want thinner deliverables, AI-ready formats, faster cycles, fewer options-papers and more structured summaries they can feed to their own tools. The majority — moving slowly — still expect the traditional form, and get confused when it does not arrive. And a meaningful subset — the AI-averse — are actively asking that AI be kept out of the engagement entirely, often with attestation, especially in regulated industries, government and confidentiality-sensitive work.
Why the split matters operationally
A firm that runs one operating mode for all three groups is producing the wrong shape of work for two of them. The AI-forward client gets material that does not survive their consumption pattern. The AI-averse client gets material whose authorship they cannot verify. The moving-slowly client is served well by accident and will be poorly served as they move.
The cost of that misalignment is subtle but compounding. It shows up as declining conversion on AI-forward prospects whose expectations the firm does not match, as quiet discomfort from AI-averse clients whose requirements the firm is not meeting, and as an inability to articulate what the firm’s own position is when asked.
How to apply the heuristic
The classification is doable from memory. Go through the top twenty or thirty accounts and mark each one as AI-forward, moving slowly, or AI-averse. The output tells you two things.
It tells you which specific clients need a different form of deliverable from you than they are currently getting — this week, not in a quarter. And it tells you whether the firm currently has one operating mode for everyone, or whether it can run two or three deliberately. Firms that drift — treating the split as a future problem and applying a single mode meanwhile — will accumulate mismatch across the book.
This is related to but distinct from Passive AI adoption is an implicit policy choice: that heuristic is about the firm’s own internal decisions by default; this one is about the firm failing to respond to clients whose decisions have already been made.