Will Customers Expect You to Pass AI Savings On to Them?

By Victor Hoang · Updated Oct 10, 2026

Yes, and the expectation shows up before the savings do. Once a buyer knows your team used AI, the question in their head is simple: if it took you a fifth of the time, why am I paying for all of it? You can ignore that question, or you can answer it with a pricing model that doesn't depend on how long the work took.

Our view is that any firm selling time, effort or headcount-shaped packages is exposed. The firms that stay in control price on the outcome delivered and the expertise needed to deliver it, and they make that move before customers do the repricing for them. This piece lays out the evidence that buyers are already pushing, a way to decide which of your offers are exposed, and a sequence for the next two quarters.

Key Facts

• In a survey of 657 in-house legal professionals across 30 countries (ACC and Everlaw, October 2025), 50% expect lower outside counsel costs and 61% plan to push for change in how legal services are delivered and priced.
• In the same survey, 59% reported no generative AI savings from their law firms yet, and only about a quarter (24%) were satisfied with how outside counsel adopt the tools for cost effectiveness.
• In a February 2026 survey of more than 130 senior lawyers, 73% of those who knew whether their firms use AI believed the firm keeps all or most of the savings.
• In a 2025 CLOC and Harbor survey, 26% of in-house respondents expected decreased spending on law firms in 2026.
• In a 2026 survey of 1,816 professionals, 78% said AI-enabled quality improvements are essential, yet 6% believe most providers deliver them, and 32% expect to reconsider provider relationships within 12 months.

Why the Expectation Is Hard to Avoid

Customers don't need a study to reason about this. They know AI drafts, summarizes and analyzes faster than people. They've used it themselves. And the moment a vendor says "we use AI," a buyer has a ready-made argument in a negotiation: the cost to serve me dropped, so the price should too.

The research on professional services shows this logic hardening into procurement behavior, even while the actual savings lag. The clearest data comes from corporate legal departments, because they buy a lot of outside time and survey well. In the ACC and Everlaw research, half of respondents expect lower outside counsel costs, and a majority (61%) plan to push for change in how legal services are priced. Yet 59% said they had seen no savings from their law firms so far.

Read those two numbers together. Expectation is running ahead of evidence. That gap is the window you have. A buyer who hasn't yet seen a discount is a buyer who is still looking for one.

The suspicion cuts the same way when firms do use AI quietly. In the Juro survey, most in-house lawyers didn't know whether their firms were using AI at all, and among those who did, 73% believed the firm was pocketing most of the gain. That's a perception, not billing data, and the sample is small. But it tells you what buyers assume when you stay silent: that you captured the margin.

And the threat isn't only a discount. In the CLOC and Harbor survey, 26% expected to spend less on law firms in 2026. The Thomson Reuters Future of Professionals research adds a retention angle: 32% of respondents said they would be reconsidering provider relationships within 12 months. That survey is about quality and implementation gaps, not price specifically, so read it as a sign that clients are actively reassessing, not as proof they want discounts.

One caution on scope. Almost all the hard survey data comes from legal services, where hourly billing makes the link between time and price explicit. The same logic applies to agencies, consultancies, software implementation partners, outsourced finance, recruiting and any business that sells hours. But we have less measured evidence there, so treat the legal numbers as the clearest early signal, not a universal rate.

Why Effort-Based Pricing Is the Real Exposure

The problem isn't AI. It's what you chose to charge for.

Here is an illustrative example, with made-up numbers. A firm bills a market analysis at 20 hours. AI-assisted, the same analysis takes 5 hours of human time. If the price is hours times rate, revenue on that deliverable falls 75% the day AI becomes routine, whether or not the customer ever asks for a discount. If the buyer does ask, you have no principled reason to refuse, because your price rested on effort and effort just collapsed.

Now price the same analysis as "a decision-ready market entry recommendation, delivered in a week, with a named senior reviewer." The buyer is paying for a result and for the judgment that stands behind it. The 15 hours you saved are yours to reinvest in more clients or deeper work. You haven't hidden anything, because the price never claimed to measure hours.

This is the same shift that our professional services piece describes from the staffing side, and it connects to a long-standing argument in billable hour versus value pricing. AI didn't create the weakness in time-based pricing. It made the weakness impossible to ignore.

Buyers already lean this direction. Thomson Reuters reports that in its 2024 State of the Corporate Law Department research, nearly half of general counsel cited an increasing preference for alternative fee arrangements. Customers who want predictability are often happy to pay a fixed price, provided it feels fair relative to the value.

A Framework: Which of Your Offers Are Exposed?

Sort every offer you sell on two questions. First, how much of the price is explained by hours or headcount? Second, how easily can the buyer judge the result for themselves?

Two lessons come out of the table. The more a price can be reverse-engineered into hours, the faster a buyer can reprice it. And the offers that hold their price are the ones where the customer is buying judgment, accountability or a result they can't easily produce themselves. That is the same logic behind why AI shifts the hiring premium to judgment: the premium moves from producing the artifact to standing behind it.

What Outcome-Based Pricing Looks Like in Practice

"Price on outcomes" can sound like a slogan, so here is what it takes.

Define the deliverable by what the buyer can do with it. Not "40 hours of analysis" but "a board-ready recommendation on which three markets to enter." If the buyer can't state what they'd do differently with the output, the offer isn't ready.

Fix the scope, not the effort. A fixed price needs clear boundaries: what's included, how many revision rounds, what triggers a change order.

Package repeatable work. If you deliver the same thing often, and AI makes delivery cheap, productize it with a standard price. Service productization is where AI-assisted margin is easiest to keep, because buyers compare a clear product to alternatives rather than auditing your hours.

Tie part of the fee to results where you can measure them. Not everything suits success fees, but some work does. The tradeoffs between metering usage and charging for results are laid out in usage versus outcome-based pricing for AI, and the same trade-offs apply to services.

Be able to explain where AI fits. You don't need to hide AI, and you shouldn't. Say plainly that it speeds up drafting and analysis, and that the price reflects the result and the senior judgment applied to it. A buyer who understands what they're paying for is much less likely to demand an hours-based discount.

What You Should Not Do

Don't cut prices in advance. A pre-emptive discount, offered before anyone asks, teaches customers that your price was padded. Raise the value you deliver before you lower the number.

Don't bury AI use. If a customer discovers it later, the story becomes "they kept the savings and didn't tell us." That is the exact perception the Juro survey describes.

Don't sell a fixed fee with a hidden hours model underneath. If you still manage the work by hours internally, that's fine, but don't let the customer feel they are paying for a stopwatch.

And don't assume every buyer reprices. Some will never ask, and some will pay for speed alone. The point is to be ready for the ones who do, and to avoid being the vendor whose only defense is "that's our rate."

Where Savings Should Go Instead

Passing savings on isn't the only alternative to keeping them. There are four places the gain can go, and which you choose is a strategy decision. We explore the tradeoff directly in whether AI gains become margin or growth.

Most firms will blend these. What matters is that the price conversation doesn't depend on the answer being "lower price." Choosing outcome pricing keeps all four options open. For the finance view on whether these investments earn their return, see how a CFO should judge AI investment returns.

What to Do Next

For a CEO, COO or managing partner, a two-quarter sequence works.

1. List every offer and tag its price anchor. Mark which are hours, headcount, or outcome. Most firms discover that more than half of revenue sits in the exposed rows of the table above.
2. Pick two exposed offers and redesign them. Choose repeatable ones. Write the outcome in the customer's words, set a fixed scope, and define what changes require a new quote.
3. Prepare your answer to the discount question. Sales and account leads should have a short, honest response to "you used AI, so why isn't this cheaper?" that points to the outcome, the accountable expert and the guarantee.
4. Pilot with new deals before touching renewals. New customers have no old price to compare against. Learn what they accept, then bring the model to existing accounts.
5. Review internal cost data separately from price. Keep tracking time internally so you know your true margin per deliverable, but stop using it as the basis of the quote.
6. Take the plan to your board. The shift changes revenue shape, so align on the risk first. A no-hype board framework for AI workforce investment helps structure that discussion.

One caveat to hold throughout. The survey evidence is concentrated in legal services, relies on perceptions and expectations rather than measured billing changes, and comes from different years and samples. The direction is consistent. The exact size of the pressure in your market is something you'll only learn from your own deals.

Learn More

• AI Workforce Transformation in Professional Services
• Billable Hour vs. Value Pricing
• Service Productization
• Usage vs. Outcome-Based Pricing for AI
• How a CFO Should Judge AI Investment Returns