AI Cut Your Billable Hours. Don't Let It Cut Your Revenue.

If your practice bills by the hour, AI hands you an awkward problem. Every time your team gets faster, the invoice gets smaller.

We worked with one professional services team on a single task they did constantly. It took about 60 minutes. Once the workflow was rebuilt around AI, it took about three. Across a year, that was roughly 1,300 hours back, worth around $250,000 in revenue.

Billed by the hour, those 1,300 hours are a pay cut. Priced differently, they’re the best thing to happen to the practice in a decade. Same tool, same people. The pricing model decides which story you end up telling.

Getting Faster Is A Pay Cut

Hourly billing has a built-in flaw that AI drags into the light. It pays you for effort. The client wants an outcome. For years those two lined up well enough that nobody worried about it.

They don’t line up any more. A first draft that took a junior four hours now takes twenty minutes and a careful review. A reconciliation that took a morning takes a coffee. If you keep billing the old hours, you’re charging for time that didn’t happen. If you bill the new hours, you’ve just cut your own revenue for doing a better job.

Neither of those is a strategy.

Clients read the same headlines you do. Bloomberg Law reported back in March that the big US accounting firms were already working on new pricing models for corporate clients because AI was cutting billable hours. Your clients may not have raised it yet. In my experience the larger ones will, and the conversation goes better when you start it.

What Clients Were Paying For All Along

Nobody ever wanted to buy your hours. Think about what a client actually values when they come to you. Being right. Spotting the thing they didn’t know to ask about. Someone qualified putting their name to the advice and carrying the responsibility for it. Getting it done before the deadline so they can stop worrying.

The hour was a convenient proxy for all of that. It was easy to measure, easy to invoice and easy to defend. AI breaks the proxy. The underlying value is still there, and in a lot of practices it goes up, because the senior people finally have time to think.

How To Price It Instead

I don’t think there’s one answer. Horses for courses. Most practices will end up with a mix, sorted by the type of work.

Defined, repeatable work goes to fixed fees. Annual accounts, tax returns, standard agreements, routine compliance. You know the scope. Set the price, describe what’s included, and let efficiency flow into your margin instead of out of your invoice.

Ongoing advice suits a retainer. A monthly fee for access, regular check-ins and a set amount of proactive work. Clients get predictability. You get recurring revenue and a reason to keep in touch beyond year-end.

High-stakes judgement can be priced on value. The restructure, the deal, the advice that changes what a client’s business is worth. Price it against what it’s worth to them.

Keep the hourly rate for genuinely open-ended work, like disputes, where nobody can honestly scope it upfront.

Moving clients across takes some nerve. The mechanics are the same as any price change, which I cover in raising prices without losing customers. Start with new clients and new matters, and move existing clients at their next engagement.

Your Timesheet Is Voting Against AI

This is where a lot of firms get stuck, and it’s a people problem more than a pricing one.

If a senior associate’s bonus depends on billable utilisation, and partner distributions track hours billed, then every tool that saves time is a threat to someone’s pay. People are rational. They’ll find reasons the tool isn’t quite right for this matter. They’ll use it quietly and bill the old hours anyway. Or they’ll avoid it and look busy.

You can roll out the best AI in the country and the incentive system will quietly kill it. I’ve seen that pattern sink perfectly good trials, which I wrote about in why AI pilots fail.

Change what you measure. Margin per matter. Realisation. Client retention. Advisory revenue per client. Reward the person who did the work in a third of the time and kept the client happy. Until the targets change, the behaviour won’t.

Plan Where The Hours Go

Time that gets saved and not planned for gets eaten by email. Every time.

Before you roll anything out, decide where the freed hours should go. More clients with the same team. Proactive reviews you’ve always meant to offer. The advisory conversations partners keep saying they never have time for. Proper training for juniors, who still need to learn the craft even when AI does the first draft, a point I made in hiring juniors in the AI era.

The 1,300 hours only became $250,000 because they were pointed at something. Hours with nowhere to go just disappear.

The Judgement Stays With The Partner

None of this changes who is accountable. These tools are probability engines. They produce a very plausible draft that is usually right and occasionally wrong in ways only an expert will catch. A qualified person reviews the work and puts their name to it. That review is what the client is paying for, and it shouldn’t be rushed to protect the new margin.

Client confidentiality matters just as much. Client files belong inside tools your practice controls, under clear rules your team actually follows. If that isn’t set up yet, an AI policy is the place to start, and the case for keeping control of the tools themselves is in renting versus owning your AI.

A Fee Conversation In 2027

Picture a client meeting a year from now. The client asks about the fee. You show them a fixed price, a clear list of what’s included, and the two things you’ll flag for them this year that they didn’t think to ask about.

They don’t ask how many hours it took. They ask whether you can do the same for their other entity.

Nobody has ever thanked their accountant for taking longer.

Frequently Asked Questions