A client opens your invoice, scans to the bottom, and pauses on a number that used to feel normal. Six months ago they'd have paid it without a second look. Tonight they're thinking about the contract their nephew drafted with ChatGPT in minutes, and they're wondering what, exactly, they're paying you for.
Your clients have stopped asking whether you use AI — they assume you do, and they've already priced that assumption into what they think your work is worth. The firms that understand this will widen their margins in 2026. The ones that don't will lose ground for a year before they figure out where it went.
What this article gives you: a clear explanation of why AI is breaking the billable-hour model, and a three-step exercise you can run on a real invoice this week to reprice around it. It works whether you run a law practice, an accounting shop, or a PR agency.
Why small firms have to reprice around AI
Because AI has split the billable hour into two kinds of work with two completely different costs. Commodity work — drafting, research, reconciliation — now costs almost nothing to produce. Judgment work — strategy, risk, ethics, accountability — costs exactly what it always did, because AI can't do it. Firms that keep charging one blended rate slowly give away the half that just got cheap, and undercharge for the half that's now their entire value.
That's the underlying case, and here's why it can't wait.
The trap you're already in
Every small firm is walking into the same squeeze. Clients expect AI to make you faster and cheaper. Meanwhile your own costs are going up — software licenses, data security, training, and the senior hours it takes to check what the AI produced before it goes out the door. Expectations pull your price down while reality pushes your costs up, and you're caught in the middle.
This isn't a fringe trend you can wait out. Organization-wide AI adoption across professional services nearly doubled to 40% in 2026, up from 22% the year before, and for the first time most individual professionals report using generative tools in their actual work. The legal market is already feeling it in pricing: the analysts behind the 2026 Report on the State of the US Legal Market expect AI-driven workflows to put real downward pressure on rates for routine matters over time. When a longtime client finally asks, "couldn't AI just do this?", the question is a fair one. The firms that get hurt are the ones holding rates flat while quietly pocketing the speed. Clients always do that math eventually.
The billable hour was always two things
The billable hour was never one thing. It was two very different jobs bundled under a single rate, and nobody had a reason to separate them — until now.
Commodity work is the production layer: research, first drafts, reconciliation, formatting, summaries, media lists. It's necessary, it's time-consuming, and it's exactly what AI now turns out at near-zero marginal cost.
Judgment work is everything the production layer is for: deciding what the research means, choosing the strategy, spotting the risk, drawing the ethical line, and standing behind the recommendation when it matters. A model can draft the memo. It cannot tell your client which of two bad options to live with, and it cannot be accountable when the call goes sideways.
For decades, bundling these together cost you nothing. AI ends that arrangement. It collapses the price of one half and leaves the other half untouched. The bundle is coming apart on its own — your only choice is whether you manage the split or let a client manage it for you.
The same story in three industries
This isn't limited to one kind of firm. The pattern repeats with almost identical shape across professional services.
Firm type | Commodity work AI now absorbs | Judgment work that stays premium |
|---|---|---|
Accounting | Tax research, reconciliation, return prep, summaries | Planning strategy, risk calls, client advisory, audit judgment |
Legal | Document review, first-draft drafting, research | Case strategy, negotiation, ethical judgment, the courtroom |
PR / agencies | Media-list building, first-draft releases, reporting | Narrative strategy, relationships, crisis judgment, the pitch |
Read it column by column and the lesson is the same everywhere: the left side is losing its cost, the right side is keeping its value. Any firm in this table can run the same exercise and reach the same conclusion.
The repricing playbook
The fix is mechanical, and a first pass takes about an hour.
Step one — tag your invoice. Pull a recent client bill and mark every line as either commodity or judgment. Don't agonize; first instinct is usually right. Most owners are quietly unsettled by how much of the bill lands in the commodity column — and that's the point.
Step two — move commodity work off the clock. Convert it to a flat fee, fold it into a package, or give it away as a free on-ramp that wins the engagement. You cannot defend a premium hourly rate for work the market already treats as nearly free, and trying to will cost you the relationship before it costs you the margin.
Step three — reprice and rename the judgment. Raise the price on interpretation, strategy, and accountability, and give it its own named line on the engagement letter. "Document preparation, 6 hrs" invites a comparison to a chatbot. "Transaction risk review and strategy" does not. None of this shrinks your revenue. It moves your revenue onto the work no competitor and no model can replicate.
The three steps, at a glance:
Tag every invoice line as commodity or judgment work.
Move commodity work to flat fees, bundles, or a free on-ramp.
Raise and rename judgment work as its own line on the engagement letter.
What it looks like on a real invoice
Here's a simple hypothetical to make the math concrete. Picture a small PR agency billing a client $4,000 a month at a blended $200/hour. Tag the lines and the split is stark: roughly 13 of the 20 hours are media-list building, first-draft releases, and reporting — commodity work AI now does in a fraction of the time. The other 7 hours are message strategy, a reporter relationship built over years, and a judgment call about when not to comment. Under the old model, the client paid $200 for both and resented half of it.
Rebuilt: the production layer becomes a flat $1,200 "content and coordination" package — honestly priced for AI-era effort, and an easy yes. The judgment layer is renamed "Strategy and Media Counsel" and priced at $2,800. Same client, similar total, completely different conversation. The client stops mentally comparing you to software, because you've stopped billing like software.
A friend at an outdoor-products company told me about a version of this earlier this year. The brand was launching a new product and wanted a long, feature-stuffed announcement. The right move was the opposite: one plain sentence about what the product actually did for the customer, with nothing that needed defending. Her client said afterward it was the best answer they could have given — simple, honest, done. The sentence itself took a minute to write. Knowing it was the only thing worth saying is exactly the kind of judgment a client should pay a premium for — and no model on the market would have told them to lead with less.
Why this grows your firm instead of shrinking it
The reflex is to read all of this as contraction — less to bill, less to charge for. The evidence runs the other way. Adoption is accelerating across the profession, and the same Thomson Reuters research that flags pricing pressure also finds the firms adopting fastest are the ones expecting to grow, not shrink. Clients aren't trying to spend less in total. They're trying to spend it on work that's actually worth paying a human being to do.
That's the durable advantage: let the machine carry the computation. Keep the judgment, the ethics, and the relationship for yourself, and charge for them openly. Show a client exactly where AI made you faster and exactly where your judgment carried the work, and the expensive lines stop feeling like something you have to justify. The firms that unbundle on purpose protect their margins and deepen what clients actually value in them. The ones that wait will have the split forced on them by a client who already ran the numbers.
So pull the last invoice. Tag it tonight. An hour with the invoice now will show you which half of your work you've been giving away.
FAQ
What does it mean to "unbundle" the billable hour?
It means separating the two kinds of work hidden inside one rate — low-cost commodity tasks AI can do, and high-value judgment work it can't — and pricing each one differently instead of charging a single blended rate.
How do I reprice my services around AI?
Tag each invoice line as commodity or judgment work, move commodity work to flat fees or bundles, then raise and rename judgment work as its own line on your engagement letter.
Will unbundling reduce my revenue?
Usually no. It reallocates revenue toward judgment work clients can't get elsewhere. Done well, total billings tend to hold steady or rise while client friction over price drops, because you are charging openly for the work clients value most.
Does this apply to law, accounting, and PR firms equally?
Yes. The commodity-versus-judgment split shows up in nearly identical form across all three, so the same three-step exercise works for any of them.
Bob Gonsalves is a brand journalist, PR consultant, and the founder of ICBM Media, a small public relations and content agency. He has spent three decades writing, advising clients, and running a practice through wave after wave of industry change. He is AI-curious: he uses AI in his own day-to-day work to save time and money, and he is still figuring out what works and what doesn't. The Small Firm AI Playbook is where he shares that AI journey as he lives it — the experiments, the wins, and the things that flop — not as an evangelist, but as a practitioner working it out in real time.