AI is changing what clients will pay professional services firms.

The unit a firm sells stops matching what the client values. Most professional services pricing meters an input the client accepted as a proxy for expertise: an hour, a head, a document set, a placement. AI makes that input far cheaper to produce while the judgment behind it stays scarce. So a firm that keeps metering the input collects less money for the same expertise, and the gap widens every quarter its tooling improves.

Exposure tracks how much of a firm’s revenue sits on a repeatable input rather than on judgment. Construction documents run roughly 31% of an architecture firm’s fee, per the Washington State OFM A/E fee guideline, and they are the phase AI automates first. In law, 71% of legal consumers now prefer flat fees and 59% of firms offer them, per Clio’s Legal Trends Report 2025, so the hourly meter is already losing ground with buyers. The larger the repeatable share of the book, the sooner AI reaches the invoice.

Start with what the firm sells rather than which tools it buys. Every sector has a repeatable half and a judgment half, and the line between them falls in a different place in each: contract review against the deal call, construction documents against the code responsibility, sourcing against the hire decision. The repeatable half is going to get cheaper no matter what any firm does about it. The judgment half is where the fee has to sit, which means naming it as an offering and pricing it on its own.