Will AI end the billable hour?

What are people actually predicting when they say the hour is finished?

Two separate claims get bundled into one headline. The first is that clients will refuse to pay for time. The second is that time will stop being a sensible proxy for value, because the same output no longer takes the same effort. They are not the same argument and they do not move at the same speed.

The second claim is already true in pockets of most firms. A document that took a morning to assemble now takes an afternoon's worth of checking, or less. The first claim is a commercial negotiation between a firm and its clients, and negotiations move at the pace of the slower party. Plenty of clients are perfectly content with hourly billing, because it puts the risk of an unpredictable matter on them rather than on their adviser, and they know it.

So the honest answer is that hourly billing will not vanish. It will retreat, unevenly, from the work where the output is visible and repeatable, and it will stay where the work is genuinely unpredictable.

What is the strongest case that hourly billing goes?

The strongest case is not about technology at all. It is that hourly billing only holds up while the client cannot judge how long something should take. Once a buyer has seen three firms quote for the same annual accounts, the same standard lease review or the same onboarding pack, they form a price expectation. From that point the hour is a unit of accounting, not a unit of persuasion.

AI accelerates this because it pushes more work into the visible and repeatable category. Buyers with any in-house capability are now testing the same drafting tools their advisers use. They have a rough sense of what the mechanical part costs. When an invoice arrives showing six hours of drafting on something they believe took ninety minutes, the conversation that follows is uncomfortable whether or not the belief is accurate.

There is also an internal pressure. A firm that bills time is a firm whose income falls when its people get faster. That is a strange incentive to carry into a period when everyone is trying to get faster. It is the point we set out in pricing work that AI makes faster: if the only lever you have is hours, efficiency looks like a loss.

What is the strongest case that it survives?

Risk. A fixed fee is a bet that the firm understands the shape of the work before it starts. On a contested matter, a complicated corporate transaction, a planning dispute or an investigation, nobody knows the shape. The hour exists because it transfers that uncertainty to the party best placed to absorb it, which is the client who chose to pick the fight.

Second, hourly billing is deeply wired into how firms run. It sets charge-out rates, it feeds utilisation reporting, it drives partner profit shares, and in many practices it is how work in progress is valued on the balance sheet. Pulling the hour out means rebuilding the management accounts and the incentive structure with it. Firms do not do that quickly, and most should not.

Third, clients are not uniform. An insurer or a large corporate buying routine volume wants a fixed price per file. A founder in the middle of a dispute wants their adviser to do whatever the situation demands and will accept a bill for it. Both exist in the same firm, and the sensible answer is a mixed book, not a doctrine.

Which work leaves the clock first?

Follow the predictability. The work that moves off time first is the work you could describe to a new joiner as a sequence of steps, because that is the work where you can estimate cost and defend a price. In practice that means:

• Onboarding, identity and conflict checks, and the paperwork around opening a new matter or engagement.
• Recurring reporting to the same client in the same format, month after month.
• Standard-form document review where the variation between jobs is small.
• Compliance packs and evidence that must be produced whether or not anything interesting happened.

What stays on the clock is advice under uncertainty, negotiation, anything adversarial, and anything where the client keeps changing their mind. That distinction is more useful to a partner than a prediction about the end of an era, because it tells you which parts of the book to reprice and which to leave alone.

What breaks inside the firm before the billing model does?

Long before pricing changes, the measurement system starts lying to you. If a report takes two hours instead of six and the fee earner still has an hours target, the hours reappear somewhere. They land in file review, in research, or in a vaguer narrative on the timesheet. Nothing has been gained and the record is now less reliable than it was. We go into this in utilisation targets after the admin goes.

The other early casualty is the estimate. Firms quote from history: what this kind of job took last time. If the underlying process has changed and nobody has re-measured it, every quote is anchored to an old cost base. That is how a firm ends up losing fixed-fee work it could do profitably, and winning fixed-fee work it cannot.

Before any of this, most firms cannot say what their own unbillable time actually costs. The method is simple, hours lost per fee earner per week, times the number of fee earners, times the charge-out rate, times the weeks they work. The guide to calculating unbillable hours sets it out and the calculator does the arithmetic.

What should a partner do about it now?

Three things, none of which require a view on the next decade.

Measure the real cost of your repeatable work. Take one recurring job and time it properly across several instances. You cannot move to a fixed fee on a number you guessed, and you cannot defend an hourly bill you have never tested. Accurate time capture matters more in a fixed-fee world, not less, because it is how you tell profitable work from the rest.

Rebuild the predictable processes before you reprice them. A fixed fee on a messy process is just a transfer of risk onto your own margin. Fix the shape of the work first, whether that is client reporting, onboarding or the compliance evidence nobody bills for, then price from the new cost. Fee estimating gets far easier once the underlying job runs the same way twice.

Choose your mix deliberately. Decide which parts of the book you want on fixed fees and which stay on time, and say so in your engagement letters rather than negotiating it matter by matter under pressure. The same logic applies to how you buy advice yourself, which we cover in fixed fee or day rates.

The billable hour will not end with an announcement. It will shrink as a share of revenue while firms that never measured their own costs keep quoting from memory. The firms that come out ahead are the ones that know, process by process, what the work actually costs them. The audit is a short way to find out where to start.