After more than a decade at KPMG, where he served as Partner, Global Head of Legal Technology, and CTO of KPMG Law UK, James Thomas (JT) made a bold move to join Leah (formerly ContractPodAi) as their Chief AI Officer. His career path is distinctive: starting as a software engineer and IT consultant, then practicing as a corporate transactions lawyer for seven years, before pioneering legal technology initiatives at KPMG. There, he deployed GenAI tools across 4,000 legal professionals in over 80 jurisdictions, working with strategic partners including Microsoft, ServiceNow, and Thomson Reuters.
Recognized by the Financial Times as one of the world's top 10 legal technologists in 2019, JT brings a rare combination of technical expertise, legal practice experience, and enterprise-scale AI deployment knowledge. We sat down with JT to understand what drove this pivotal decision and where he sees the future of AI heading.
FII9 was extraordinary, genuinely one of the most impressive gatherings I've attended in my career. The seniority of attendees was staggering. I've never watched a panel with the CEOs of JPMorgan Chase, Goldman Sachs, KKR, Carlyle, Pershing Square, Intel, all sitting at one table answering questions in person. That alone tells you about the caliber and importance of the conversations happening there.
But beyond the star power, three things really stood out to me.
First, Saudi Arabia's approach to AI infrastructure. They're not making one-off investments, they're building a scalable AI ecosystem through long-term commitments via the Public Investment Fund, Humain, and now partnerships with Aramco. These are sovereign-backed initiatives designed to accelerate deployment and commercialization of AI across industries. It's strategic, it's patient, and it's massive in scale.
Second, the compute conversation has fundamentally shifted. Global compute is now being discussed as the next strategic resource, on par with oil, water, or energy. The CEO of Groq had a quote I loved: "AI is built on compute, the fourth civilization-level resource." Power consumption requirements are exploding as demand outstrips supply, and solving that challenge has climbed to the top of the global agenda.
Third, and this ties directly into why I joined Leah, the event reaffirmed something critical: the AI orchestration layer is going to become essential infrastructure for every organization. Everyone is talking about it, but very few have actually cracked it. We have, with Leah Agentic OS.
At FII9, I had conversations with government officials, enterprise leaders, and investors—all grappling with the same challenge. They're deploying AI tools across departments, but it's fragmented. Marketing has one AI tool, legal has another, finance is using something else, and nothing talks to each other. There's no unified governance, no orchestration, no enterprise-wide intelligence.
What became crystal clear is that agentic workflows and agentic systems have moved from conceptual to real. Governments are leaning on AI to digitize public services, improve outcomes, and deliver more value to citizens. Enterprises are doing the same across legal, procurement, finance, HR, IT - you name it.
But here's the problem: without an orchestration layer, you just have chaos. You have dozens of point solutions creating silos, security risks, and inefficiency. What Leah has built with Leah Agentic OS is that missing infrastructure, a single, unified platform where intelligent agents work together, share context, and optimize across functions.
That's not a nice-to-have anymore. It's essential. And seeing the convergence of that realization at FII9, across industries and geographies, confirmed we're building exactly what the market needs.
I enjoyed my time at KPMG immensely - it’s a fantastic firm, and it gave me so many opportunities in my career to grow and develop. I started there as a Corporate lawyer and quickly moved into a technology role. In that role, I had the privilege of building technology, designing and implementing our strategy, and deploying technology, including GenAI tools like Leah, across a global network of 4,000 legal professionals in over 80 jurisdictions. I had a great team, and together we designed and built some incredible solutions.
When I started working with Leah as a customer of the company, I always really enjoyed working with them, and the tech and their roadmap really stood out. Leah was one of the very few companies in the early days that saw the step change in AI capability and embraced GenAI rapidly when so many others were busy trying to figure out how to augment existing systems and grappling with legacy tech debt around machine learning.
As Leah’s tech and capabilities continued to develop, I realised that where I wanted to be was to be at the centre of that, helping to drive development forward - directly influencing the roadmap and strategy. When I look at what Sarvarth, Anurag, and the team have created, partnering with both of KPMG and PwC, OpenAI, Anthropic, Google, and Microsoft - it’s where the future is being built, not talked about, and I wanted to be a part of that.
Yes, and it's more than a buzzword - it's a fundamental shift in how AI operates. Let me break it down.
For years, we've had AI tools that respond when you ask them something. Think ChatGPT or Microsoft Copilot: you input a question, you get an output. That's useful, but it's reactive. Agentic AI is different. These are systems that can autonomously identify objectives, navigate complex environments, make independent decisions, and take action to achieve predetermined goals. They don't wait for the next instruction; they adapt, learn, and expand their capabilities as new data emerges.
In the legal context, imagine an AI that doesn't just draft a contract when you ask, it proactively identifies contract risks across your entire portfolio, suggests remediation strategies, monitors regulatory changes, and automatically flags provisions that need updating based on new compliance requirements. That's the shift we're talking about.
What excites me is the liberation it offers professionals, whether they're lawyers, procurement officers, or public servants. Look, I've been both a software engineer and a corporate transactions lawyer. I've lived on both sides of this equation.
Professionals are brilliant strategic thinkers, but they're drowning in repetitive, data-heavy work that doesn't require their expertise. Agentic AI handles the procedural tasks, the document review, the clause extraction, and the compliance checks, so people can focus on what they're actually trained for: judgment, strategy, and relationships.
This isn't about replacing people; it's about empowering them. At KPMG, I saw firsthand how legal teams were spending 40% of their time on contract management alone. With Leah Legal and the broader Leah Agentic OS platform, we're not just addressing that 40%, we're transforming nearly 100% of legal activities through custom legal applications, from M&A data rooms to GDPR compliance to IP portfolio management.
And that scales beyond legal. That's why Leah Agentic OS can orchestrate agents across legal, procurement, finance, HR, IT, and more. The opportunity is enormous.
I think, as a software engineer, it helps to understand what's technically possible and where the boundary between hype and reality is. As a corporate lawyer, I worked through the real-world pains of what lawyers experience day to day and understand how lawyers think and work (we’re pretty risk-averse!). I think the combination ultimately allows me to bridge the gap between what engineers build and what practitioners need.
I know from my own experience that so many solutions are built by technologists who've never practiced in the domain, or by domain experts who don't understand the technology's capabilities and limitations. My personal view is that I think you need both perspectives to build something transformative.
Absolutely valid concerns, and I think the industry has taken them seriously.
Organizations should be cautious because we're dealing with confidentiality, privilege, and sometimes high-stakes decisions. Everyone has read about the use of LLMs in court filings, which ended up hallucinating cases and decisions - it’s a real risk that I think most people are not alive to.
I think what separates enterprise-ready AI from consumer-grade AI is governance, explainability, and domain specificity. At Leah, we're not just layering AI onto legacy tools. Leah’s solutions are built AI-first, leveraging deep domain experience and enriched with domain context.
The other important factor is transparency and control. No one particularly likes a black box when it comes to decision-making - you want and need to understand how an answer was reached. We’re combining both deterministic and probabilistic methods to achieve that - mixing our Agents in with traditional “If this then that” logic and a mix of pre-built data transformation code blocks to ensure the underlying language models are fed the right information at the right time in the right format, and we can properly validate the outputs within a workflow.
Ultimately Legal professionals need to understand how the AI reached a conclusion, what data informed it, and have the ability to validate outputs, and we’re doing that - providing explainability and allowing professionals to maintain oversight while gaining efficiency.
We're moving from the era of AI assistants to the era of AI agents, and then beyond that to full agentic operating systems. Leah is already building this with Leah Agentic OS, which provides enterprises a single, unified infrastructure to build, orchestrate, and govern intelligent agents across functions.
Right now, every department in an organization is experimenting with different AI tools in silos. The future is a coordinated, enterprise-wide agentic platform where intelligent agents work together, share context, and optimize across functions. That's not five years away - it’s here now.
The market is expected to grow at over 40% annually, potentially exceeding $50 billion by 2030. But more important than the market size is the transformation it represents. We're not talking about incremental efficiency gains. We're talking about fundamentally redefining how organizations operate.
I'd say you're right to be thoughtful and cautious, but you can't afford to ignore this. The professionals who understand and can effectively leverage AI will have an enormous advantage over those who don't.
My advice is to start using the tools more in your day-to-day - understand prompt engineering, learn how to use these tools effectively, and, critically, understand their limitations and dangers. This should be taught in law schools and business schools from day one. Future-proof yourself and your practice.
The technology isn't perfect and it's evolving rapidly, but I think the trajectory is clear: AI is transforming how work gets done, and those who embrace it strategically will lead their professions. Those who resist will be left behind.
I think the future is always difficult to predict, and the pace at which things are changing is difficult for anyone, even those in the industry, to keep up with. Models are exponentially improving in capability, and integration across ecosystems is accelerating. This means organisations and individuals are going to finally be able to achieve the aim of “more with less”.
It’s a hugely exciting time - the pace of change is going to keep accelerating - the workplace we know today is going to look very different in five years' time, and it’s exciting to be at the cutting edge of that transformation.
James Thomas joined Leah as Chief AI Officer in October 2025, bringing over a decade of experience leading global legal technology initiatives at KPMG and a unique background combining software engineering and corporate law expertise. He was recognized by the Financial Times as one of the world's top 10 Legal Engineers / Legal Technologists.
