Skip to content
upshift
The Three Shifts

AI is repricing your firm. The question is which model survives.

Billable hours, retainers, project fees, outcome pricing. AI is changing all four. What protects margin is the same in each.

No amount of tooling fixes a pricing model that’s working against the firm. When the price a client pays is tied to the time a team spends, and AI cuts that time, the math runs the wrong way no matter how good the technology gets.

The pricing model is where the AI conversation actually lives. Not at the tool layer. Not at the implementation layer. At the line on the invoice.

The firms that win this don’t buy more AI. They change what their clients pay them for.

Each model is taking different damage from AI. The fix in every case is the same: tie the price to what the client actually buys, not your time and not your tools.

What’s already in the field.

Roughly a quarter of McKinsey’s global fees now run on outcome-based structures, per UK managing partner Michael Birshan in November 2025.

The four pricing models, and what AI does to each.

Which of the four carries most of your revenue today?

AI works for you

Result-priced

The outcome

What AI does

Faster results, better margin. Your incentive aligns with the client’s.

What’s working

Name the result. Price it. Deliver it.

Scope-priced

The project fee

What AI does

Clients expect fees to drop to reflect AI savings.

What’s working

Charge a premium for AI-enabled deliverables that weren’t possible before.

Access-priced

The retainer

What AI does

Fewer touchpoints. Clients question what the retainer is buying.

What’s working

Sell named outputs, not access.

What AI does

Compresses hours and revenue. Clients now expect 20% less.

What’s working

Sell deliverables, not time.

AI works against you

Time-priced

The billable hour

Sell deliverables, not time.

You charge per hour worked. The hour is the unit, time is the meter.

What AI does

AI compresses the hours, which compresses the revenue. The faster your team gets, the less you bill. That’s not theoretical. Clients are already telling firms they want the AI savings. PwC’s Chief AI Officer Dan Priest put it this way: “Clients would hear us talking about using AI and say, ‘We want our fair share of those efficiencies.’” If your revenue is tied to hours, that share comes out of your margin.

What’s working

Start with a revenue audit, not a pricing overhaul. Sort your work: which of it will clients pay more for as AI makes it faster, and which less? The cheaper half is execution. The premium half is judgment, bundled invisibly into the same hour. Pull it out and price it on what being wrong would cost the client. That number comes from them, not your timesheet.

Access-priced

The retainer

Sell named outputs, not access.

You charge a fixed monthly fee for ongoing access: counsel, advisory, on-call expertise.

What AI does

Retainers priced on access feel softer to clients every quarter that AI absorbs more of the hands-on work. Clients see fewer touchpoints, fewer billable activities, and start asking what the retainer is buying. Retainers that survive don’t price hours of availability. They price named outputs the client recognizes as work AI didn’t do.

What’s working

Convert the retainer from access to named outputs the client recognizes: a quarterly strategy review, a decision-tree update, a board-ready memo. The pricing question becomes “what did you decide differently because of this work?” But naming the outputs isn’t enough. Write down what the retainer doesn’t cover, in plain terms. A retainer that names deliverables but leaves the edges open quietly turns back into unlimited access at a fixed fee, and the margin goes with it.

Scope-priced

The project fee

Charge a premium for AI-enabled deliverables that weren’t possible before.

You charge a fixed price for a defined scope of work, regardless of how many hours go into it.

What AI does

Project fees were already scope-based, not time-based, so they look safe. They’re not. Clients see AI as the engine of efficiency and expect the fee to come down to reflect it. The trap is negotiating yourself into 30% lower fees and absorbing the entire AI productivity gain as client savings. That’s the same compression as hourly billing, one step removed.

What’s working

Hold the fee where the deliverable is genuinely faster, and build the AI-enabled work that wasn’t possible before into a separate offering. A 48-hour contract analysis isn’t a faster version of the old review. It’s a different thing the client couldn’t buy last year. The trap is pricing it off what it now costs you to produce, which is almost nothing. Price it on what the speed and the certainty are worth to the client when the deal is on the clock.

Result-priced

The outcome

Name the result. Price it. Deliver it.

You charge for the result, not the work. Price reflects what the outcome is worth to the client, not what it cost you to deliver.

What AI does

Outcome pricing aligns your incentive with the client’s. AI makes you faster, the result lands sooner, your margin improves instead of compressing. Roughly a quarter of McKinsey’s global fees now come from outcome-based pricing, per UK managing partner Michael Birshan in November 2025. CB Insights briefed more than 25 AI agent vendors and found that while 54% still use subscription licensing, nearly a quarter already offer outcome- or performance-based pricing.

What’s working

Name a result you can stand behind, and pick one inside your control. “We’ll cut your monthly close from a week to 24 hours” is priceable because you own it. “We’ll win you the deal” isn’t; too much of that depends on the client. Before you start, agree the number you’re measuring against today and what counts as hitting it. Skip that and the result arrives as an argument over who caused it instead of an invoice.

Calculator

Your firm’s number.

You’ve read the ladder. Two minutes to see what hourly billing costs you and outcome pricing pays you.

See your number

What ties them together.

Every model that survives ties the price to what the client actually buys. Not your time. Not your tools. The result they came to you for.

TSIA’s Professional Services 2.0 framework, published in March 2026, says it directly: firms must shift from selling “hours, headcount, activities” to selling “time-to-value, business outcomes, measurable impact.” Their research director, Bo DiMuccio, put it this way: if your revenue still depends on hours or headcount, AI will continue to erode your margins.

The work is to name the result, price it, and stand behind it. That’s what protects margin in every one of the four models.

One thing has to move with the price. If your partners are still measured and paid on hours, repricing a single offering won’t hold; the first quarter it costs someone billables, the old model comes back. Changing what you charge means changing what you reward for delivering it.

The Pricing Ladder runs from hours to outcomes, one of the Three Shifts. The Workshop names and prices 2–3 of those offerings, each defined tightly enough to put in front of named clients for a real yes or no. Your team or an implementation partner builds the ones that earn it. For the longer argument behind outcome pricing, read Outcome Pricing Captures the Upside. For how it lands in your sector, see the industries page.

Questions

AI compresses the time required to deliver professional work. If your revenue model is tied to time, your revenue compresses with it. The fix in every case is the same: tie what you charge to what the client actually buys, not your time and not your tools. Each of the four pricing models (billable hours, retainers, project fees, outcome pricing) takes different damage, so applying it looks different model to model.

Outcome pricing charges for the result, not the work. Price reflects what the outcome is worth to the client, not what it cost the firm to deliver. AI makes the firm faster, the result lands sooner, and the firm’s margin improves instead of compressing. Roughly a quarter of McKinsey’s global fees now come from outcome-based pricing, per UK managing partner Michael Birshan in November 2025. CB Insights found that nearly a quarter of AI agent vendors already offer outcome- or performance-based pricing.

Fixed fees price a defined scope of work at a set price, regardless of hours. Outcome pricing prices the result the client wants, regardless of scope. The fixed fee asks “what does this work cost?” Outcome pricing asks “what is this result worth?” The first ties price to delivery cost. The second ties price to client value.

Pick one offering, not all of them. Convert your most defined work (contract review, monthly close, audit, fixed-scope analysis) to a deliverable fee first. Once that lands, the next conversion is easier. Trying to switch the whole book at once produces partner disagreement and stalled implementation.

The fee can stay the same. What’s priced has to change. Replace “hours of availability” with named outputs the client recognizes as work AI didn’t do: a quarterly strategy review, a decision-tree update, a board-ready memo. If your retainer line item describes hours, AI is going to commoditize it.

Negotiating yourself into 30% lower fees because the work is faster. Clients see AI making the work cheaper and expect the savings to flow back. The trap is conceding the discount without changing what you’re selling. If you can deliver an offering that wasn’t possible before, charge a premium for the new one. Don’t discount the old one.

Document the result well enough that the next client buys it without negotiation. Outcome pricing works once. It scales when you can show the result you produced for the last client and price the next engagement off that. Keep what you can name. Drop what you can’t.

Ready to figure out what you sell next?

Thirty minutes with Shawn Yeager. No pitch. Talk through what the Workshop looks like for your team.

Book a conversation

Or see where you stand