Fixed fee is a systems problem.

Fixed-fee pricing is the single clearest external marker of an AI-native practice, and the most commonly attempted without the work underneath it. The pattern is predictable: a firm decides that transparent pricing is a competitive advantage, publishes numbers derived from instinct, absorbs three matters that run long, watches margin evaporate, and concludes that fixed fees do not work in its practice area.

The pricing model was not the problem. The absence of duration data was. The billable hour trained the profession to record time rather than manage it, and the distinction only becomes visible when you try to quote in advance. Recording tells you what you billed. Managing tells you how long the work actually takes, how much it varies, and why. Only the second lets you name a number without gambling.

Run it the other way and pricing becomes close to mechanical. If your standard lease review has taken between four and seven hours across the last twenty instances, with a median of five, you can price it, buffer it, and be right most of the time. The pricing question dissolves once the measurement question is answered.

But measurement has a prerequisite of its own, which is why this is a systems problem rather than a pricing one. You cannot measure duration meaningfully against a matter as a whole - eleven hours on the Sharma matter tells you nothing reusable. You measure against named actions. Which means the action library has to exist first. Which means the matter has to be defined first. The dependency chain runs all the way back to the beginning of the sequence.

This is the part that makes fixed pricing so informative as a signal. A firm that has committed to it publicly has, necessarily, done the naming and systematising work - because the alternative is losing money in a way that shows up quarterly. It is a claim that cannot be made cheaply, which is exactly what makes it worth reading.

There is a second-order effect worth noting. Once a firm prices flat, its incentives on automation invert. Under hourly billing, an automated hour is deleted revenue. Under a fixed fee, the same automated hour is margin. Firms that make this switch before adopting AI find that every subsequent technology decision gets easier, because the organisation is finally rewarded for the thing the technology actually does.