AI Agents for Law Firms: AI Is Saving Hours. The Billable Hour Isn't Letting Firms Keep Them.

Law firms have stopped asking whether to use AI. The harder question now is what to do with the hours it saves, in a business that has always been paid by the hour.

PwC's 35th annual Law Firm Survey 2026, published in October and covering the UK's Top 100 firms, puts numbers on that tension. Adoption has jumped. Productivity gains are showing up. But very little of it has turned into money, and firms are more worried about AI's effect on their business than they were a year ago.

Two Thirds of Top Firms Have Now Implemented AI

The adoption curve moved quickly this year. According to PwC:

• 66% of Top 100 firms have now implemented AI, up from 43% last year.
• More than 75% of Top 50 firms have invested in specialist legal AI assistants.
• Dedicated AI teams are now normal. Top 10 firms typically have more than 20 full-time people on AI, while Top 51-100 firms have between one and five.
• 52% of firms seeking transformational change plan to use AI to reshape service delivery or turn services into products, compared with 42% focused on efficiency.

PwC's own summary is blunt: AI "has moved to the centre of the sector's strategic agenda."

The Hours Are Being Saved. The Money Is Not Showing Up.

The gap between using AI and getting paid for it is the survey's most striking finding.

• Only 8% of Top 100 firms have turned AI productivity gains into meaningful commercial returns.
• Just 7% have saved more than 10% of chargeable hours. Most firms report savings of less than 5%, although every Top 10 respondent reported savings of 6-10%.
• Firms now estimate that around 18% of chargeable hours could eventually be saved through AI, up from 16% last year.
• Where firms do see value, it shows up in fee earner productivity (25%), quality control (23%) and responsiveness to clients (19%).

PwC names the reason directly. The gains are held back by gaps in data, training, measurement and operational integration, "all of which are compounded by a deeply human reluctance to step away from the familiar comfort of the billable hour."

That is the core problem. If an associate finishes a first draft in two hours instead of five, a firm that bills by the hour has just earned less for the same work, unless it changes how it prices, staffs and measures that work.

Firms Are More Worried, Not Less

Confidence in AI's commercial effect has fallen sharply even as adoption rose:

• 78% of Top 100 firms now expect AI to have an adverse effect on their business, up from 66% in 2025 and 31% in 2023.
• Concern that clients will automate work and rely less on outside counsel rose from 53% to 68%.
• Concern about the speed of technological change rose from 73% to 84%.
• Only 30% are concerned about disruptive competitive models, which PwC calls a possible "strategic blind spot."

PwC notes this pessimism comes "despite no signs of productivity declining." The worry is about the business model, not the technology.

Policies Are Ready. The Plumbing Isn't.

Governance is the one area where firms are well ahead. 70% of Top 100 firms report well-established AI policies and controls. But fewer than a quarter have well-established:

• training (23%)
• data quality and access (13%)
• benefit measurement (5%)

In other words, firms have written the rules for AI faster than they have built what AI needs to do useful work: clean, reachable data, people who know how to use it, and a way to tell whether it is paying off.

The workforce data hints at where this is heading. Paralegal and legal executive headcount fell across every firm banding, by 7.5% to 14.5%. Chargeable hours are declining in many senior roles. And only 14% of Top 10 firms have applied strategic workforce planning firm-wide.

Sign up for our newsletter

Get the latest updates, product launches, and helpful tips delivered straight to your inbox.

Email addressSubscribe

Where AI Agents Help a Law Firm Most

Most legal AI so far has been assistant-style: a lawyer asks, the tool answers. The value PwC says firms are missing sits in the work around the legal work, and that is where AI agents, which can carry out multi-step tasks across a firm's systems, are a better fit.

• Matter intake and conflicts. Gather new-matter details from email, run them against the firm's records, and prepare the conflict check for a person to approve.
• Billing and lock-up. PwC reports average lock-up rose to 147 days at Top 10 firms. An agent can chase time entries, prepare draft bills, flag matters drifting from budget, and send reminders on unpaid invoices.
• Fixed-fee and value pricing support. Track actual effort against agreed fees, matter by matter, so partners can price AI-assisted work with real data instead of guesses.
• Measuring the benefit. Only 5% of firms have well-established benefit measurement. Agents can log time saved per task and per matter, giving firms the evidence they need to change pricing.
• Client updates. Draft routine status notes from matter records for a lawyer to review and send.

None of this replaces legal judgment. It removes the administrative work that keeps lawyers from billing, and it creates the data firms need to decide how to price a faster service.

Where Workmate Fits

PwC's findings point to a firm-wide problem, not a tool problem: governance is in place, but data access, training and measurement are not. Workmate builds custom AI agents around a firm's own workflows and connects them to the systems it already uses, from email to practice management. Our team sets each agent's routines and permissions, then launches it inside Microsoft Teams, Slack, or email, where lawyers and business services staff already work.

For firms that guard client data closely, the controls matter. Admins decide which apps each agent can reach, each Workmate has its own separate computer and private file storage, and credentials are stored in a vault that can require a person's approval before use. Workmate is also SOC 2 Type II audited.

What This Means for Law Firms Going Forward

The firms pulling ahead in PwC's survey are not the ones with the most AI tools. They are the ones turning saved hours into a different business: sharper pricing, leaner back offices, and services clients will pay for in new ways.

For everyone else, the next year is about closing the gap between the 70% with AI policies and the 5% that can measure what AI is worth. That starts with the unglamorous work of intake, billing and reporting, which is exactly where AI agents earn their keep.

Frequently Asked Questions

How many law firms use AI in 2026?

PwC's Law Firm Survey 2026 found 66% of the UK's Top 100 law firms have implemented AI, up from 43% the year before. More than 75% of Top 50 firms have invested in specialist legal AI assistants.

Is AI making law firms more money?

Not yet for most. Only 8% of Top 100 firms have turned AI productivity gains into meaningful commercial returns, and just 7% have saved more than 10% of chargeable hours.

How much time could AI save lawyers?

Top 100 firms estimate around 18% of chargeable hours could eventually be saved through AI, up from 16% last year. Most firms report current savings below 5%.

Why does the billable hour slow down AI at law firms?

When AI makes work faster, a firm billing by the hour earns less for the same task unless it changes how it prices and staffs work. PwC cites reluctance to step away from the billable hour as one reason gains are not turning into returns.