Ari Kaplan recently spoke with Sean Fitzpatrick, the CEO of LexisNexis' legal business, where he leads strategy, growth, product innovation and operations.

They discussed the value of proprietary content in an AI era, perspectives on the potential shift in how AI is priced, how customers benefit from partnerships with different AI providers, and evolving customer expectations for legal AI.

Ari Kaplan: Tell us about your background and your role at LexisNexis.

Sean Fitzpatrick: I’ve been at LexisNexis for 21 years, and I am currently the CEO of our legal business, responsible for our global strategy, product development and bringing the product to market across all aspects of the legal business.

Ari Kaplan: How is AI affecting the way your team serves its law firm and legal department customers?

Sean Fitzpatrick: Everything’s accelerating. We just finished the first half for the legal business. Our underlying revenue growth was 10%, an acceleration from the first half of last year. That equates to just under $80 million in incremental revenue in the first half of this year that we didn’t have last year. If you annualize that, it’s roughly $150 to $160 million in incremental revenue. We did that in the first half of ‘26. We did something similar in ‘25. We did something similar to that in 2024. It’s becoming a much more meaningful part of our overall business. If you look at where that growth came from in the first half of the year, 90% of our new business was from our generative AI products. So it’s having a big impact in the marketplace. Customers are seeing value from these products, and they’re voting with their wallets.

Ari Kaplan: What value are you seeing in proprietary content in an AI era?

Sean Fitzpatrick is the CEO of LexisNexis’ legal business, where he leads strategy, growth, product innovation and operations.

Sean Fitzpatrick: The models are only as good as the content they’re built on, and we have a significant trust issue in legal. We now have about 1,800 documented instances of attorneys bringing materials into court that cite cases that don’t exist. They’re hallucinations from AI, and that’s not slowing down. It’s happened to some of the largest firms in the U.K. and in the U.S. If it can happen to them, it can happen to anyone. So there’s an important trust layer now that everyone’s very focused on. And it’s not just about efficiency. Everyone knows you can get efficiency from using these tools, but it’s about grounding them in trusted content and adding a verification layer, so you can go back and validate that something’s actually correct. The models produce content that sounds really confident, but what our attorneys need is to be confident in those answers. We can help them be confident, not just by grounding it in our authoritative legal material but also by giving them access to that. We’ll give them a Shepard’s trust signal, so they know it’s a real case. They can click on it and see the actual case. They can read the headnotes to find out what the points of law are. They can read the caselaw summary to learn about the case. They can review the Shepard’s report to see if it’s still good law.

Ari Kaplan: What are customers saying about the potential shift in AI pricing?

Sean Fitzpatrick: One of the things we’re hearing is a lot of concern about moving to a token-based pricing model, and that’s creating a lot of concerns for our customers because they want to have certainty, or at least fairly good certainty, about what they’re going to be paying for their solutions. We’re not moving to a token-based model. We’re sticking with our subscription model, and we’re signing customers into three- and five-year deals. Over the course of that contract period, they’ll have price certainty. We’ve had a multimodel approach from the very beginning, and the combination of that approach and our proprietary content set allows us to get answers to questions using more distilled models that are really focused because we know exactly where the answer is. If somebody has a question about regulatory information, for example, we can point that model right to our regulatory database and quickly find the answers, and that’s a very efficient way to use tokens. If you’re trying to find out what the latest regulation is and you’re searching every corner of the internet to find that out, that’s a very token-intensive process. It’s a competitive advantage that we have, and it’s the reason that we’re not moving to token-based pricing. In fact, token costs represent less than 1% of RELX revenue.

Ari Kaplan: How do your customers benefit from your partnerships with various AI providers, and how do those partnerships align with LexisNexis’s broader strategy?

Sean Fitzpatrick: We have partnerships with OpenAI and Anthropic, and those partnerships give us advanced access to the tools before they’re released, so we can understand how they work and how they might benefit our customers. We also have access to engineering talent within those organizations, who help us better understand those models and how to apply them to legal use cases. That’s accelerating our product development process, and our customers benefit from that. We recently signed an agreement with Luminance. Their AI product is called Lumi. If you’re a Lexis+ with Protégé customer and a Luminance customer, you’ll be able to ask a question that requires our legal research information to answer within Lumi. It will take that question, encrypt it and send it via API to Lexis+ with Protégé. Protégé will generate the answer. That encrypted answer will come back, and it will be available to the Luminance user within that Lumi format. So we’re also trying to make the AI products that our customers buy available in more places.

Ari Kaplan: How do you balance investing in new AI capabilities with strengthening your core business?

Sean Fitzpatrick: The legal information analytics and decision tools business we’ve been in for a long time is benefiting from AI because it’s now an AI-powered ecosystem of information analytics and decision tools, and you get to your answers more quickly and more efficiently. Beyond that, there’s a massive market that’s starting to develop, and analysts estimate it will be five times the size of the total addressable market we have for legal technology today. We’re expanding into that section of the market, as well. It represents a huge growth opportunity. I talked about the results we’re seeing. A lot of it’s from going into that white space.

Ari Kaplan: Given the importance of technology in the legal industry today, what skills are you seeing lawyers develop that give them a competitive advantage?

Sean Fitzpatrick: Certainly, understanding a little bit about large language models, how they work and how you can efficiently find the information you’re looking for is an important skill. So much of this knowledge is transferred just by working with others, so there’s a collaborative element to it, as well. One thing I’m hearing from law firms is that soft skills are becoming increasingly important, so there’s an additional emphasis on developing those skills in firms now.

Ari Kaplan: How do you see law firms and corporate legal departments investing in these tools to drive growth within their organizations?

Sean Fitzpatrick: We serve both law firms and corporations, as well as government entities. In corporations, in-house counsel tends to be really stretched in terms of the amount of work they can take on. Using AI tools, they’re able to complete more work than they were able to before. They’re able to do it faster, so turnaround times are quicker, and they’re able to do it with a greater level of confidence, so we see internal teams really benefiting from being able to get to legal issues that they weren’t able to address before. With law firms, we’re seeing lots of different ways that they’re benefiting from it. They’re certainly benefiting from the efficiency component, but there are also all kinds of new legal issues that are being generated by these new products, which is creating more legal work.

Ari Kaplan: What’s next on LexisNexis’s product roadmap?

Sean Fitzpatrick: We’re really working with our customers to drive the future product roadmap, and one of the things they’re very interested in is skills. So you take a particular task you want to complete and put in a set of rules for completing it, exactly how you want it done, the content sets you want to draw on and so on. That’s a great way to get a consistent result across your entire firm and to use your AI assets in a very efficient way. We’ve just introduced skills, and our customers are really grabbing onto it, so we see that as a big growth driver going forward.

Ari Kaplan: How do you see customer expectations for legal AI evolving?

Sean Fitzpatrick: One of the key ways it’s evolving is the growing importance of trust. We’ve seen too many instances of law firms getting caught with hallucinations in their filings, and it has a massive impact on those firms’ reputations. We’re seeing fines and penalties increase. We used to talk about $5,000 fines. I just read about a $110,000 fine. I heard about an attorney getting suspended for six months. So the stakes are starting to go up, but would you use a law firm that had just been sanctioned? You have lots of choices when it comes to law firms, so reputation really matters, and that’s a big shift we’re seeing. Folks are moving toward systems they know they can trust, so they do not end up in that situation.

Listen to the complete interview at Reinventing Professionals.

Ari Kaplan regularly interviews leaders in the legal industry and in the broader professional services community to share perspective, highlight transformative change and introduce new technology at his blog and on Apple Podcasts.