Who Should Own AI at a Law Firm?

At a Glance

• AI usually arrives at a law firm as a procurement question. That routes it to IT, and the strategic question never gets asked.
• IT should own security, integration and the cost of running systems. Fee earner productivity and pricing need an owner in firm leadership.
• In-house clients are spending less and concentrating work on fewer firms, so how a firm uses AI is turning into a question about the bill.
• Firm leadership needs written answers on margin, client expectation and accountability before AI spend goes beyond a pilot.
• A managed service provider keeps systems running. Redesigning how legal work gets done is a separate engagement with a separate owner.

Ask who owns AI at a law firm and the answer usually comes with a nod down the corridor, towards IT.

On paper, that makes sense. AI shows up as software, software has to be bought, and IT buys software. So a law firm AI strategy often begins life as a vendor evaluation: is it secure, does it connect to our document management system, what does it cost per seat.

Those are fair questions, and the firm needs the answers. Whether AI changes the firm's economics depends on a different set: pricing, staffing and what clients will pay for. The people who can answer those are rarely in the room when the tool gets chosen.

Why Law Firm AI Strategy Keeps Landing on the IT Desk

The route is predictable. A practice group sees a demo, or a client mentions a tool, and someone asks IT to "have a look". IT then runs the process it runs for every new system: security review, data residency, integration, licence cost.

By the end, the firm knows whether it can run the tool safely. What the tool should change is still an open question.

Cost pressure makes that routing stick. The Law Society's Financial Benchmarking Survey 2026, which covers 121 firms in England and Wales with combined fee income above £1.2bn, names the ongoing rise in IT costs as the main factor behind an increase in non-salary overheads. It gives two causes: consolidation in the software supplier market, and increasing AI spend.

So AI now sits inside the overhead line partners already watch most closely. When that line gets reviewed, AI is judged like any other overhead, by what it costs. What it earns rarely makes it onto the page. We looked at how to separate those two arguments in law firm IT budget planning for 2027.

Uneven adoption pushes the decision further down the organisation. The ABA's 2024 Legal Technology Survey Report found that 30% of responding lawyers were using AI tools, rising to 46% at firms with 100 or more lawyers (ABA Law Practice Tech Report). Where use is patchy, AI looks like a collection of tools to support, and supporting tools is IT's job.

What IT Should Own in a Law Firm AI Strategy, and What It Should Not

IT still has a big part to play. A firm that rolls out AI without its IT team is asking for a data incident (and the head of IT will be the one explaining it to the insurer).

IT is the right owner for:

• Security, access controls, and where client data is stored and processed
• Integration with the systems the firm already runs, from practice management to document management
• Support, reliability and the run cost of each tool once it's live
• Vendor due diligence and contract terms

The commercial side is where IT can't carry the decision alone. Should a matter type move to a fixed fee once drafting time falls? Do an associate's saved hours go into more matters, faster turnaround or fewer write-offs? Does the client get a cheaper bill, or does the firm keep the margin? Those are partnership decisions, and they shape revenue for years.

When IT is the only owner, AI gets measured on uptime and logins. Licences get bought, adoption gets counted, and nobody can point to a change in the P&L. That's usually the moment a partner asks whether the pilot was worth it, and the room goes quiet. We traced the same pattern on the legal operations side in why AI hasn't fixed legal operations yet.

The Three Questions Behind a Law Firm AI Strategy

Firm leadership doesn't need to pick tools. It does need answers to three questions, written down, before AI spend goes beyond a pilot.

1. What is AI doing to our margin? If a task that took 6 hours now takes 2, someone has to decide where the other 4 go. On hourly billing they vanish from the invoice unless volume rises. On a fixed fee they become margin. Each practice group will land somewhere different, and that call should be made on purpose.

2. What is it doing to client expectations? Clients read the same headlines partners do. Some will expect lower bills, some faster turnaround, and some will want to know exactly what AI touched on their matter. The firm needs a position it can state the same way every time, before a panel review forces one.

3. Who is accountable? A named person at partner or COO level who answers for results across margin, client experience and risk. IT stays accountable for the systems. This person owns what the systems are for.

The first question leads straight into pricing, which we covered in how top law firms are rethinking pricing. The third has a practical side too: a named owner can only stand behind AI output if the right checks sit in the workflow, and AI governance for law firms sets out what a CTO has to build.

What Clients are Now Asking About Law Firm AI Strategy

In-house legal teams are working with less. The ACC Law Department Management Benchmarking Report 2026, run with Major, Lindsey & Africa across 576 legal departments in 45 countries, puts total legal spend at 0.43% of company revenue, a six-year low. The median in-house lawyer now supports 367 employees, up from 300.

Those teams still send most of their outside spend to law firms: 93% of it, according to the same report. And they concentrate it, with around 80% of work often sitting with a handful of firms.

Put those numbers together and the pressure on panel firms is plain. A legal department with a tighter budget and a short panel list has every reason to ask its firms how they use AI, and what that means for the bill.

In our conversations with UK firms, the client questions tend to fall into four groups:

• Is AI used on our matters, and for which tasks?
• Where does our data go when it is?
• If the work takes less time, how does that show up in what we pay?
• Who checks the output before it reaches us?

IT can answer the second question. The other three need someone who knows the firm's position on pricing and quality. Without one agreed answer, different partners will give the same client different responses, and that's a harder conversation to repair later.

Turning a Law Firm AI Strategy into Something with an Owner and a Number

AI strategy papers are easy to write and easy to ignore. The version that survives a budget review attaches an owner and a number to each piece of work.

A practical sequence looks like this:

1. Pick 2 workflows where time and cost are already measured, such as billing narrative review or first-pass document review. (Bill review is where many firms start; we compared the options in build or buy an AI billing assistant.)
2. Record how long each one takes today, and what it costs.
3. Name a partner or practice lead who owns the result, with IT owning the systems underneath.
4. Agree the number that would justify carrying the spend into next year.
5. Decide in advance what happens to the time saved: lower price, more capacity or kept margin.

Then run it for a quarter and report the number, good or bad.

Many firms already have a managed service provider (MSP) and ask whether that partner can take this on. An MSP keeps systems patched, supported and running, and good ones do it well. Redesigning how a billing review or a disclosure exercise gets done is a separate engagement. It needs people who understand legal workflows, can rebuild the process around the tool, and are measured on the business result. Plenty of firms keep their MSP and bring in AI implementation capability alongside it; they're two contracts with two different jobs. We set out how those contracts differ in Cognitive Outsourcing™ vs managed services.

That second job is the work 3Rive does. Our AI services team builds legal AI tooling that runs inside the firm's own tenant, so client data stays in the firm's environment. Through Tech Advisory, AI Readiness in Five Days assesses infrastructure, process and governance, and ends with a costed 90-day plan a management board can approve.

If your firm's AI decisions still sit with IT alone, an AI readiness assessment is a practical way to put the three questions in front of the partnership with evidence attached.