When AI shrinks the distance between starting a project and finishing it, the hour stops measuring what it used to.
The billable hour was never really the product. It was a proxy. A way to price expertise and risk using the one unit of accounting that both sides could measure and agree on. As long as complex work reliably took a predictable number of labor hours, that proxy held up.
AI breaks the correlation it depends on.
What happens when delivery time drops but the value doesn’t?
Firms still billing for time instead of outcomes will increasingly charge less for more.
Imagine a fully functional data pipeline was built in a matter of days using agentic delivery, with AI handling much of the build work a team would normally spend weeks on. The client’s outcome didn’t change. They got the pipeline they needed, to the same standard. What changed was the number of hours behind it.
Priced by the hour, that project bills for a fraction of what a comparable one used to, even though the value delivered can be the same or greater. The cost to deliver dropped. The client’s benefit didn’t. Something in that gap may need to be change.
What does it look like to price the outcome instead of the hour?
One response is to stop billing by the hour and price against the outcome instead. That doesn’t have to mean a single flat number. It might be a fixed fee for a defined deliverable, milestone payments tied to specific checkpoints, or a retainer priced against maintaining a result rather than the hours spent maintaining it.
Take a project like the one mentioned above, where a working data pipeline was built in days instead of weeks. Priced hourly, that project’s value to the firm shrinks the moment AI compresses the timeline. Priced against the delivered pipeline instead, the fee holds regardless of whether it took three days or three weeks, because the firm gets paid for the outcome, not the clock.
The tradeoff is real. Hourly billing tolerates loose scope because time absorbs the ambiguity. Outcome-based pricing moves that risk onto the firm, so “done” has to be defined tightly before the work starts, not judged after the fact. It also asks more of the client conversation upfront: hours and dollars are numbers people already understand, and an outcome has to be described in enough detail to be priced at all.
What does it look like to keep the hour, but change what happens inside it?
Not every firm is moving away from hourly or project-based billing, and that’s a reasonable choice too. Some are keeping that structure and using AI to change what happens inside it: covering more of the client’s problem within the same scope, or bringing a level of quality and thoroughness to the work that wouldn’t have been realistic in the same number of hours before. The client isn’t paying less for the same thing. They’re getting more for the same price, which is its own way of turning AI’s efficiency into value rather than into a lower bill.
This path lets a firm capture AI’s efficiency gains inside a structure clients already understand, without renegotiating how engagements get scoped or priced.
Is there another way to think about what AI is actually contributing?
A third idea: translate what AI does into an equivalent number of human hours. An AI agent might do in minutes what would take a skilled person days, so the AI’s own clock time tells you almost nothing about how much work actually got done. Given what the AI actually processed and produced on an engagement, how long would that same output have taken a person to read, reason through, and write from scratch? That’s the number this approach is after.
Right now, that figure is mostly used internally, giving a firm a way to see, engagement by engagement, how much capacity AI is actually freeing up. But the interesting question is where it goes from here: once that kind of estimate is trusted enough, could it be something a firm could eventually price or credit work against directly?
Is there a “right” answer yet?
Not yet, and probably not just one. None of these paths rule out the others, and a firm doesn’t have to pick a single one. What’s changing is that the old default, billing for the exact number of clock hours worked, is becoming more complex in an AI-accelerated engagement. Choices have to be made. The firms figuring this out first are the ones treating that as a real strategic question, rather than assuming the hour will keep working the way it always has.
Ready to price for what your AI-powered team really delivers?
Innovative Solutions helps organizations rethink how AI-accelerated work gets scoped and priced. Reach out to talk through what that could look like for your business.
