AI is changing the economics of legal work. That much is clear. What is less clear is how law firms should price work when the inputs, effort and delivery model are also shifting.

There is a lot of noise around whether AI is saving lawyers time. It is an important question, although it is difficult to answer properly without knowing exactly what lawyers put into AI tools, what comes out, and how that changes the real work done on a matter. That level of analysis is important, but it is not the most immediate commercial issue for most firms.

The more pressing question is simpler. If firms want to give clients greater certainty, why are so many still pricing matters through broad ranges and loose assumptions?

The awkward gap between AI value and pricing

BigHand’s 2026 Pricing and Budgeting Trends Analysis captures the tension well. The report finds that “AI is changing how legal work gets delivered faster than firms are changing how they price it.” It also shows that operational progress is moving ahead of pricing maturity: “reduced manual/repetitive work has risen from 28% to 31% this year”, while “only 29% of firms have changed how they price AI-augmented work, and only 20% report any improvement in pricing accuracy or consistency.”

This is the hard bit. Delivering value and capturing value are different disciplines. A firm may use AI to complete certain work faster, then still struggle to decide who should benefit economically. The firm could keep the efficiency gain. The client could receive a lower fee. The fee could stay steady because the client is receiving better certainty or a stronger outcome. Each choice can be reasonable if it is deliberate and backed by evidence.

At the moment, the market has not settled. Asked how they currently reflect AI-driven efficiencies in pricing, firms have no dominant approach: “30% are maintaining existing pricing while banking the efficiency gain internally, 30% are still evaluating what to do, and 25% are offering discounts when clients permit AI use.” That should concentrate minds. Nobody has won the argument yet.

Solving data problems solves pricing problems

My view is that the industry needs to get a proper grip on matter economics. That means understanding the work inside a matter at a level that is useful for pricing and resourcing conversations. Timesheets are usually treated as a billing record. They are also one of the richest sources of evidence a firm has about how work was actually delivered.

If that data can be read carefully, structured consistently and grouped into meaningful workstreams, firms can identify which past matters are genuinely similar to the one being scoped. That is a very different starting point from asking a partner to estimate from memory or asking a pricing team to build a model from imperfect assumptions.

This is where AI becomes practically valuable in pricing. The useful application is the ability to turn messy matter history into a clearer view of likely effort, work phases, staffing patterns and budget risk. That gives partners a firmer basis for quoting. It also gives clients a better experience because the number they hear is less likely to drift every week.

It matters even more for AFAs, when many lawyers are naturally attracted to them because they can align well with how clients want to buy legal services. The problem is risk. Under many AFAs, the firm carries more of the economic exposure. Without a defensible view of how long tasks are likely to take and which resources are needed, the fee can become uneconomical very quickly.

As demand for AFAs grows, that pressure lands heavily on pricing teams. Firms can keep adding analyst capacity, or they can use technology to make good pricing intelligence available earlier in the process. The latter only works if the system is grounded in trusted delivery data. Otherwise, it is a faster route to the same uncertainty.

Don’t be shy of the commercial conversation

The client side of this is also moving. BigHand’s report says “53% of firms say clients are willing to pay the same fee where AI improves value or outcomes, while only 26% believe clients would pay a premium and 17% expect AI to reduce what clients pay.” That 53% figure is not a safe hiding place. There is a difference between a position tested with evidence and a position that feels calm because the market has not yet forced the issue.

The report also finds that the leading barrier preventing firms from communicating technology use more proactively is that “35% cite partner discomfort discussing AI with clients as their biggest obstacle, ahead of regulatory concerns (31%).” I do not think that means partners are afraid of AI. I think it means they do not yet have the data, pricing policy or governance needed to make the conversation commercially safe.

That is why this is becoming one of the most pressing topics in the industry. Ayora, which is part of the BigHand suite, addresses this problem. Its data enrichment layer helps turn unstructured matter data into structured, contextual intelligence, giving firms a more reliable foundation for commercial insight and AI-supported pricing decisions.

The aim is to give firms better intelligence before those critical pricing decisions are made.

Price with evidence to set the pace

The slightly uncomfortable conclusion is that firms celebrating AI efficiency may still be commercially behind. Measuring matter economics, deciding their AI pricing position, and defending that position in client conversations will only become more mandated.

Firms still have time to define their position. To do that, leaders need to decide how AI-enabled work should be priced, put governance around those decisions and build pricing conversations on evidence rather than instinct.

The full BigHand 2026 Pricing and Budgeting Trends Analysis explores these issues in more detail, including how firms are responding to AI transparency, client expectations and the positive pressure for more accurate matter pricing.

Download the report to see the data behind the shift.