Death of the billable hour? Keystone Law's Maurice Tunney on AI and the platform law firm
Keystone Law's Maurice Tunney with David Woolstencroft on People in Legal.
Many law firm technology leaders inherit decades of systems patched together for different departments and ways of working. Maurice Tunney, Director of Technology and Innovation at Keystone Law, got something rarer: the chance to build from scratch for a firm that works completely differently.
Keystone is a listed business, not a partnership. Its roughly 500 partner-level lawyers are effectively self-employed. They keep 75% of what they bill, and Keystone keeps 25% to provide the platform, brand, marketing, compliance, PI insurance and back-office support. When Maurice joined almost 12 years ago, there were about 100 lawyers.
In this episode of People in Legal, Maurice explains how that model shapes every technology decision, from time recording to generative AI. Here are the insights that stood out.
1. Treat your lawyers as your clients
At Keystone, the internal customer really is the customer. "Our clients are our lawyers," Maurice says, and the firm aims to give them a white-glove service so they can focus on their own clients rather than compliance and anti-money laundering.
The commercial design reinforces collaboration rather than silos. Lawyers earn 15% of any work they refer to a colleague, which removes the traditional partner's reason to guard contacts. The result: by Maurice's estimate, around a third of last year's billings were internally referred. That flywheel now lets Keystone build practices for strong lawyers who arrive without a client book, including those coming from in-house roles.
The lesson: culture follows incentives. If you want lawyers to share work and knowledge, pay them to.
2. Cloud-first from day one pays compounding dividends
With no legacy estate to protect, Keystone went cloud-only. The stack has four legs:
Microsoft Office suite for Word, Excel, PowerPoint and Outlook
NetDocuments for document management
Teams for messaging and video
Keyed In, Keystone's own platform, for everything else
Keyed In handles time recording, client and matter creation, digital engagement letters, write-offs, billing and financial reporting. It also includes a collaborative feed and a map showing where lawyers are based, so someone heading to Bristol can find a local employment lawyer for coffee.
Because the systems are process-driven, a team of four onboarded in a couple of hours the week before we spoke.
The lesson: scalability is a design decision. Firms that standardise processes in the platform can grow without adding friction.
3. Build for how lawyers actually work
Maurice found many off-the-shelf tools capable but rigid. Keystone's in-house developers build around lawyer behaviour instead. Some lawyers run a live timer. Some type their time at the end of the day. Some still use paper and do a monthly bulk import. Keyed In supports all three.
The latest release, a pop-out timer that stays visible while lawyers work in a document or email, came straight from lawyer feedback.
The lesson: adoption is earned, not mandated. Flexibility in how a tool is used often matters more than the feature list.
4. Match AI to ability, not hype
Maurice describes AI as "terrifying and exciting in equal measure." Keystone's approach starts with an honest look at each lawyer's comfort with technology, then matches tools to that level, backed by training and guidance.
The layers look like this:
Keybot: secure ChatGPT access built on open-sourced guardrails from Shawn Curran, with model choice, document upload, projects and a shared prompt library
NetDocuments ND Assistant: improving fast, and every user inherits new functionality as it ships
A manual search assistant: ask a plain-English question about Keystone's 350-page operating manual and get an answer with references, contained to that document
A client compliance agent: six months in development with an agency, it handles awkward compliance questions with clients in a secure, contained way
Opt-in CoCounsel licences: Keystone negotiated a deal with Thomson Reuters and co-funds licences, with lawyers paying a share so they have skin in the game
Access to Thomson Reuters content lawyers already use was the differentiator in choosing CoCounsel, and the choice stays under review as new tools emerge.
The lesson: a portfolio of AI tools, matched to user capability, beats a single big-bang rollout.
5. Lead with the use case, not the tool
One lawyer told Maurice he estimates his practice grew by 60–70% in a year, which he credits largely to Keybot for drafting, analysis and chronologies, all under close supervision. At the same conference, another lawyer said there was simply too much new technology to keep up with.
Maurice's answer is to show practical wins rather than hand over tools. His example: an employment lawyer uploads a client's staff handbook, compares it against their own gold standard and sends back a redline, perhaps even as a free service. That builds the stickiness every firm wants, so when a dispute arises, the client calls them.
The lesson: busy lawyers sink when given tools without context. Show them one job AI does better, and adoption follows.
6. AI is raising the bar on both sides
Clients now expect more, but they are also meeting new problems. Some Keystone clients are fielding sharper complaints from their own customers, drafted with ChatGPT. Smaller clients increasingly arrive with AI-generated documents and ask lawyers to simply sign them off, missing the nuance that creates risk.
The lesson: AI makes some work easier and other work harder. Professional judgement becomes more valuable, not less.
7. The billable hour's time may finally be up
Maurice has heard about the death of the billable hour for his 25 years in legal. This time, he thinks it is real: when AI slashes the time some work takes, charging for time stops making sense. The shift will be to outcome-based pricing.
That puts data at the centre. Firms will need to know how long similar work took, who did it and what the outcome was, so they can price the next job with confidence. As we discussed in the episode, time recording starts to become job costing.
The lesson: start building your matter data now. You cannot price outcomes you have never measured.
8. Expect specialists, not certainty
Asked where things will be in a year, Maurice is candid: he has no idea, and he is sceptical of anyone who claims to know. Three tools Keystone is now implementing did not exist a year ago. He expects more niche products that solve one problem well, as vendors realise they cannot win the land grab at the top. Meanwhile, his own developers now ship new functionality up to ten times faster using AI coding tools.
The constant, he says, is the job itself: working out what lawyers really need, because they know where they want to get to but not always how.
What legal leaders should take from this
Design incentives that reward sharing, and collaboration follows
Build process into the platform so growth does not add friction
Give lawyers flexibility in how they use tools
Tier AI by capability, and fund power tools jointly so users commit
Sell AI through practical use cases, not training sessions
Capture matter data now to prepare for outcome-based pricing
Watch the full conversation
People in Legal is where legal leaders and legal technology meet. Want your leaders in front of our audience at the next event? Partner with us or email team@peopleinlegal.com.
Quotes lightly edited for clarity.
Comments