AI in Legal Services: What the 2026 Data Says About Cost
AI is changing legal work. Is it changing what you pay?
A couple of years back, most corporate legal departments were still shying away from AI. A 2024 survey from Deloitte Legal found that 76% of legal firms had no AI adoption at all. However, when Deloitte ran the same survey in 2026, that figure had dropped to just 2%.
But whether legal teams are adopting AI is only part of the story. The shift has knock-on effects on pricing, market expectations, turnaround times, recruitment and more. The biggest mismatch between what firms offer and what clients expect comes down to pricing. The people buying legal services expect AI to change what they pay and how they pay it. So far, the way most legal work is priced hasn't moved much.
Below, we look at what the data says, why that gap exists, and what it means if you run a company without its own legal department.
The findings in this article come from Deloitte Legal's The AI imperative: Reshaping of the legal industry, which surveyed 121 senior in-house legal leaders, and Thomson Reuters' Future of Professionals Report 2026, which surveyed 736 law firm professionals across 46 countries, plus 203 people working in corporate legal departments.
The transition from novelty to industry standard
In-house legal teams have adopted AI faster than almost anyone expected. In Deloitte's survey, seven in ten legal departments have moved past experimenting and into real deployment. One in ten say that AI is fully embedded in daily work across the whole team.
And with adoption comes spend. 79% of legal departments increased their AI budget this year, and those that did raised it by 67% on average.
Today, adoption is strongest where the work is heavy on documents. Contracts and commercial work leads, and around half of respondents already use AI at scale for reviewing, summarizing and comparing documents. On average, legal leaders expect AI to save or automate 28% of their department's work within two to three years. For contracts and commercial work, the estimate is 34%.
Even the teams furthest ahead still want a person accountable for the answer. Legal advice generated by AI and given straight to the business is the least adopted use case, with 64% of legal departments not using it at all. Only 24% have a defined quality assurance framework for their AI use.
Clients expect AI to change their bill, but few have seen it yet
Both reports find the same gap, looking at it from opposite sides.
Thomson Reuters asked in-house legal professionals what they expect from the law firms they hire. 70% say it's very important or essential that firms show efficiency gains from AI, such as faster delivery or lower cost. Only 6% say most of their firms actually deliver this.
The same in-house group was asked about pricing. 71% expect firms to change their commercial model, meaning how they charge, as AI use grows. On the law firm side, 28% say they have changed their pricing structure in response to AI.
Deloitte found the same pattern among General Counsels. 85% believe AI will change how law firms price their work. Yet 58% say their external providers rarely or never bring up AI benefits with them, and just 4% have directly experienced a benefit from a provider's use of AI.
While efficiency is showing up inside legal teams, it hasn't reached the invoice yet.
Why the savings aren't reaching clients
Law firms are investing in AI, and many feel the pressure to go further. Thomson Reuters found that 38% of law firm professionals already face financial pressure to move faster on AI. The reason those gains haven't shown up for clients comes down to the billing model.
Most legal work is still billed by the hour. Deloitte puts the share of hourly-rate work at 72% today. When AI cuts the time a task takes, an hourly model can turn time saved into lost revenue for the firm doing the work.
Deloitte's report points out that as AI adoption increases, the billable hour will become significantly harder to defend. It also highlights that there's no simple fix. Moving to fixed fees means rethinking the whole business model, not just the price list, and every client relationship is different.
This conundrum leaves firms with a real choice to make about where AI's value goes. According to Thomson Reuters, only 3% of law firms describe their direction as rebuilding around AI from the ground up. That's the path that includes outcome-based pricing, subscription-style support and working as an extension of the client's team. Most firms are using AI to strengthen the model they already have.
Clients want more than a lower price
It's easy to read this data as clients simply wanting cheaper legal work, but the Thomson Reuters data is more specific. The in-house professionals who expect a change in pricing fall into three groups, and the largest one isn't asking for a discount.
Predictability is the most common request, and one of the least delivered. Higher fees for better outcomes is the only row where firms deliver more than clients ask for.
Deloitte's data reinforces this trend, showing that legal leaders expect the share of work billed by the hour to fall from 72% to 44% within two to three years, and from 74% to 48% in Europe, the Middle East and Africa.
Most clients aren't asking lawyers to absorb the whole saving. 42% think savings from AI should be split equally between the provider and the client. This fair-split mentality shows that clients aren't only pushing for savings. They want a price they can see before the work starts, and predictability throughout the engagement.
What this means if you don't have a legal department
Neither report surveys companies without in-house lawyers. Deloitte's respondents run legal departments with anywhere from under 25 to over 200 people. Thomson Reuters surveyed law firms and the in-house teams that hire them. A 60-person scaleup with no lawyer on staff isn't in either sample, and from that position the story looks different.
Large companies are spending serious money to capture AI's benefits for themselves. They're hiring legal engineers, testing AI agents and moving work in-house. 61% of legal departments are already experimenting with agentic AI, and about two thirds say bringing work in from external firms is a priority. Deloitte notes that external legal support typically costs three to five times more than in-house resources, and some General Counsels have set targets to cut external legal spend by 20 to 40% over the next three years.
While insourcing more legal work might make sense for large corporations, the leap is too big for fast-growing companies without a legal department. There's no legal team to train, no budget for legal tech platforms and not enough volume to justify one. The obvious shortcut is a general-purpose AI chatbot. But even legal departments that have embraced AI mostly aren't using it to give advice with no lawyer involved, and fewer than a quarter have a framework for checking its output.
Where Frank fits
Put the two reports side by side and the gap is clear. Clients want to know what legal help will cost before the work starts, and they expect some of what AI saves to reach them. Companies without their own legal team are the furthest from getting either.
Frank was designed to fill that gap. We're an embedded legal counsel service for growing companies that don't have a legal team of their own. We use AI in how we work, but every answer has a real legal advisor standing behind it.
We're built on the model only a small share of law firms have moved to: working as an extension of your team and pricing for outcomes, not hours. When you work with Frank, you pay a flat fee, so you know your legal cost up front.
Every legal provider should be able to tell you what happens to the time AI saves them. We think you deserve a frank answer. If you'd like to hear ours, talk to us.
Frequently asked questions
How is AI used in legal services today?
Today, AI is mostly used for work built around documents: reviewing and summarizing contracts, drafting first versions, legal research and finding information in large sets of files. Deloitte found that around half of corporate legal departments already use AI at scale for document review. Advice given with no lawyer involved is still rare.
Will AI make legal services cheaper?
It makes much legal work faster, and clients expect that to show up in price. 85% of General Counsels in Deloitte's survey believe AI will change how law firms price their work. Whether you pay less depends on how your provider charges. Under hourly billing, a faster lawyer doesn't automatically mean a lower bill.
Is the billable hour going away?
Not overnight, but legal leaders expect it to shrink a lot. In Deloitte's survey, the share of work billed by the hour is expected to fall from 72% today to 44% within two to three years.
Can I use an AI chatbot instead of a lawyer?
For general background, a chatbot can help. For decisions with real consequences, such as employment contracts, shareholder agreements or data protection, you need someone accountable for the advice. Even well-resourced legal departments are cautious here: 64% don't use AI to give advice directly to the business.
How much legal work can AI automate?
In-house legal leaders estimate AI could save or automate an average of 28% of their department's work over the next two to three years, according to Deloitte. The estimate is highest for contracts and commercial work at 34%, and lowest for specialist areas like competition law at 19%.
Sources
Written by
Meagan Leber
meagan@frank.legal
