For Iain Telford, AI Success Depends on People, Not Just Tools

I think the most useful thing about my background is that none of it was accidental. I was a law student heading for traditional practice until a student innovation competition changed my mind. I didn't win, but reaching the final convinced me there was an opportunity to build a unique career within the legal profession.

I went to the University of Edinburgh and obtained an LLM in Innovation, Technology, and the Law, then made a series of deliberate moves through RegTech, law firms, and the vendor side, each chosen to get closer to the role I have now. I also took the New York Bar, because credibility matters when you're asking lawyers to rethink how they work. It's harder to dismiss the innovation conversation when the person leading it is qualified to do the job differently.

The reason I mention any of that is because I think it shapes how I work. I'm not someone who was handed an innovation portfolio and asked to figure out what to do with it. I came to this with a view, built over my career, of what good looks like and what doesn't. A lot of what gets called "innovation" in this industry is neither — and being able to tell the difference is a crucial part of the job.

Manchester United and Hibernian, which between them in recent times have given me an advanced education in patience. Live music — indie rock mostly, but I'll go to almost anything. And travelling or spending time outdoors whenever I can. If there's a thread, it's that I'm drawn to things where the outcome isn't guaranteed and the people involved actually care. Not a bad description of how I work either.

Joining Maples Group last year. Not because of anything I did, but because of what I inherited. Innovation is one of the firm's three strategic pillars — it's not a discretionary initiative, it's how Maples defines itself. The firm had already made some genuinely innovative calls before I arrived, and the instinct in a new innovation role is to arrive with a plan and start changing things. I learned very quickly to do the opposite. The most useful thing a new innovation leader can do is assume the people who came before them knew what they were doing, and earn the right to change things by understanding them first.

The conversations have changed. For most of my career, innovation in legal services was a defensive conversation: efficiency, margins, catching up. The last number of years have flipped it. Now we're talking about what new work becomes possible and what the firm of the next decade looks like.

Having watched a lot of technology get oversold and underdelivered, I'd add this: it feels like we're now in a wave where reality has actually kept up with the hype. It helps that Maples is investing in a new Irish headquarters in Dublin — and when the physical environment, technology, and strategic direction are all moving together, the momentum becomes self-reinforcing.

Maples is celebrating twenty years in Ireland this year, and that same instinct — commit early and build for the long term — is exactly how we approached AI. The original Harvey call predates me. I was very lucky to step into my role alongside a very forward-looking Partnership. We were Harvey's sixth global client and the first in Ireland, selected when the platform was still relatively unknown. That decision has aged extremely well.

Harvey is now a core piece of a much wider innovation portfolio, and that's been my focus since joining: building out from a strong base, broadening adoption, and connecting it to everything else we're doing. Picking the right platform is maybe 20% of the job. The other 80% is what you do with it once it's in the building.

We have seen genuine adoption across all of our practice areas, but two patterns have surprised me. The easy assumption is that GenAI is a junior-lawyer tool. Our experience has demonstrated that some of our most senior people are among the most ambitious users — they see it as leverage for the judgment calls only they can make, not as a replacement for junior capacity.

The other surprise is how effectively use cases have travelled between offices on their own. Something starts in one jurisdiction and shows up in another within days, with limited central push. That's very difficult to engineer — when it happens, you mostly try not to get in the way.

There isn't a single approach; there's a system, and the parts reinforce each other. Champions networks, practice-led training, leadership sponsorship, embedding in workflows, a shared library of prompts. The one I'd actually emphasise is the shared library. After time on the vendor side, I know how much energy goes into generic enablement. What actually moves adoption inside a firm is much more specific: a colleague in your practice, in your jurisdiction, used Harvey for this exact task, and here's how. Peer evidence beats training every time.

Workflows and agentic engineering are what interest me most and where we've seen the greatest impact. What excites me isn't any single use case, it's the pattern.

In some practices, we've moved past the stage where Harvey is a tool people reach for when they have a spare moment. It's now embedded in how work gets structured and delivered. The shift is from "I used Harvey on this task," to "this workflow was designed with AI as a core component from the start." I see a future where AI-enabled delivery is no longer a support function or side project, but a workflow engineering capability embedded within practice groups – built in tandem between innovation and legal teams, owned by the people doing the work, and evolving with every matter that runs through it.

I'd be suspicious of any innovation leader who gives you a single number, and even more suspicious of one who only measures the GenAI piece in isolation. Across the innovation portfolio, we track time saved, depth of usage across practices and offices, quality of work product, client feedback, and the broader commercial picture. What matters is looking at them together, and over time — month-on-month trajectory tells you more than any single snapshot.

Each one in isolation can be misleading: Time saved without quality is a problem. Quality without adoption is a pilot. Adoption without commercial impact is expensive enthusiasm. You need them all together, or it's very difficult to know.

I have two, and both are slightly uncomfortable to say out loud. First, the platform matters — get that wrong and nothing else lands. The firms that get the most out of GenAI aren't just the ones with the best tooling. They're the ones that take the human side seriously enough to properly invest in it. And you learn faster by starting than by planning. Maples was an early Harvey client because the Partnership backed a call before the playbook existed. The firms who wait for the picture to clarify are the ones who'll be furthest behind in two years.

Practical advice for anyone starting out: pick a technology partner you can build a real relationship with, not just a product. Pick a few high-value use cases. Find the people inside your firm who actually want to do this work. Measure what matters rather than what's easier to count. And don't wait for perfect — it's crucial to build momentum and build success incrementally.

Invest heavily in building the right team to be the architects of your journey. I'm a big believer in the concept of "Purple People" — people who sit between disciplines and can translate in both directions. Not pure technologists, not pure lawyers, but people who understand enough of each world to connect them. Those are the people who turn an innovation strategy into something that actually ships. If you get the team right, most of the other problems are easier to solve.