Hi, everybody. Next up, we have Thomson Reuters. A company needs no introduction, but obviously a leading provider of business information services. Solutions include highly specialized information-enabling software and tools for legal, tax, accounting, and compliance. Very pleased to have Steve Hasker, General Manager of Global and Large Law Firms, and Gary Bisbee, Head of Investor Relations, here to provide us with an update. Welcome, gentlemen.
Thank you.
Thank you.
Steve, why don't we jump right into it? What are you seeing from large law firms in terms of how they're adopting AI and the pace of that innovation? If you have any data on kind of wallet share within these organizations and how to think about that.
Yeah. Well, maybe I'll kick off, and then Gary, if you want to provide any data around wallet share.
Yeah
I'll defer to you.
Perfect.
I see demand accelerating significantly.
Okay.
I was at an event just this past week with chairs and managing partners from some of the world's largest law firms, and they attended all of our optional sessions around roadmap, where the product's going. Even two years ago, that didn't happen. We're seeing engagement from the top of the house and all the way down. Every member of the firms now are interested in generative AI. They see it as a core part of their futures, and they see us as a core part of where they're going. We have a lot of engagement right now.
Okay.
Yeah, I don't know that we have a number to quote on wallet share. I think a helpful framework to think about in terms of the AI solutions in legal, from my perspective is, we have the traditional markets where we've been the market leader from a market share perspective in the content-enabled areas, legal research, Westlaw, and we have a major competitor, and on the transactional side of law, Practical Law.
There's a number two there. We've not seen real change in the dynamic competitively. We're innovating. Our primary competitor's innovating as well. The accelerating growth we've seen serving law firms in our legal business, I think a key driver of that is the upgrade cycle within Westlaw. So we're certainly really pleased with how we're doing in that category. There's also this new, really exciting market, which is AI workflow. By the way, in those research areas, we are innovating with AI in a big way, and that's a big driver, one of the big drivers, of the accelerating revenue. In the AI workflow space, which is largely new and incremental for us, we did not have a big revenue contribution from workflow software prior to AI that's being disrupted. The key question is how much of this new market are we winning?
Certainly, there are other competitors there, some newer AI startups. Some of the model providers are adding some capabilities. We're doing really well. The growth of CoCounsel is a big driver of our acceleration. I think we feel good about our positioning there today. I'd say our strategy in that market is clear, which is to leverage two of what we believe are really important advantages Thomson Reuters brings in the form of our authoritative content, both on the legal research and transactional side, and our deep base of subject matter experts through our 2,500 attorney editors, which are important in helping create these AI solutions, but also fine-tuning the outputs, so that we can deliver the accuracy that our customers need.
I think we feel good about that, how we're doing in that market, and are really excited about the next generation version of CoCounsel that is in beta right now, and we've talked a bit about in the last couple of earnings calls.
Can I just piggyback for a moment.
Yeah, absolutely.
off of the new category he mentioned, the AI workflow category. I think that one's a very dynamic and rapidly evolving marketplace. I think from my vantage point, there's kind of three tiers of competitors or players in that space. One is the foundation model providers.
like the Anthropic, the Google, Gemini. The other is kind of the wrappers, so just putting a wrapper around those large language models. The third tier is the one that we're moving into with our new generation of CoCounsel, which is the fiduciary-grade AI layer. That's taking like Claude Agent SDK's agentic harness and applying that natively to our content. With the agents that come forth from that, we're going to be able to do things that we believe no one else in the market's going to be able to do because those agents are going to be able to adapt in real time and provide output that is grounded in our content.
A pushback I get pretty regularly from some investors is just around the content that you guys have versus content a Harvey or somebody could go out and train on themselves. Can you give us a bit of an update on how you think about that competitive advantage for Thomson Reuters?
Yeah. Let me start and h appy to have Steve supplement. First of all, we discussed the Westlaw content advantages and why we believe we've got durable advantage there on our third quarter conference call in early November last year. I would urge anyone who really wants to dig into that to look at the slides and the transcript from that section that we discussed. I think the key takeaway from that is we believe that it would be incredibly difficult, costly, and timely for anyone to replicate anything close to Westlaw, and there's several reasons. First of all, we have 150-200 years of this content, depending on the market, and a lot of that you just can't get. There's also a lot of content that we get that isn't available digitally.
We add a significant layer of intelligence onto that content through the work that our editors do with a number of forms of editorial enhancement, then we've got software tools and other capabilities we've built up over time that allow you to make sure that you've considered everything, that it's comprehensive. 99% isn't good enough when you're thinking case law, because if the other side has a case you've missed, that could be the determining factor in the output of your case. We also have key tools to prove out accuracy. KeyCite, the citator, not only helps you find citations, but also is how you determine if a case was later overruled in part or in total. This is a lot more than just scraping some data that's publicly available.
The last point I'd make, in an AI world where you have outputs that are AI-generated, it is crucial that you validate what those outputs are before you use them in the high-stakes work that we serve customers. Westlaw has the best tools to validate the outputs and make sure that they're not hallucinated. There've been a number of cases of law firms that have used large language models and come up with outputs that included hallucinations, Westlaw is the best way to make sure you're not doing that. We just don't think that that combination of content, editorial enhancement, and significant technology tooling that we've built up is easily replicable.
We also don't really see competitors in the market today, at least not that I'm aware of, trying to replicate the same depth and breadth of content that we have.
Yeah.
The wrappers, and you mentioned one, are more kind of differentiating at the workflow level. They're trying to build out proprietary workflows, and they're also trying to differentiate on UI, and they hire legal experts to go into firms and partner with the firm to drive adoption. None of them are really trying to replicate that content advantage because they just really don't have the ability to do that.
Yeah. You've obviously seen very strong growth in the legal group, 9% organic growth. How do you think about it in terms of the different segments? You mentioned the renewal cycle, upgrade cycles. How do you think about the sustainability of that kind of 9% type growth?
Let me take a first cut at that. I get asked a lot, what's the composition of the 9%? In my former longtime sell side brain, there's three ways you could drive revenue growth. You could have pricing, you could have volume, meaning more users, more people adopting your solution, or you could have mix, which is selling a higher capability, higher priced version of a product category they already buy.
I think our legal growth that we've been delivering is a pretty good mix of the three. We've talked in the past of about 3% price, though with some of our newer AI solutions, it's worth noting we're asking for larger price increases than that in multi-year contracts due to the accelerated pace of innovation we've delivered. If you think sort of 3% price, we have a lot of that concept of mix or upgrades. That's Westlaw. That is we launched last August Westlaw Advantage with the agentic deep research capability. We've seen a tremendous start in terms of adoption or upgrades to Advantage. If you looked at, we said on the last call that for the law firm business, taking out government, we grew 11% in Q1, up from 9% in Q4 and 8%-9% through last year.
The biggest driver of that 9 to 11 is the strong upgrade cycle at Westlaw Advantage that we've seen through the first eight, nine months of the adoption of that product. We have volume, meaning more customers adopting new solutions, and that's CoCounsel, the AI assistant, which is the other key driver of the revenue acceleration we've seen in recent years and recent quarters, along with the Westlaw upgrade cycle. It's a good blend of price, volume, and mix. We believe that based on the innovation roadmaps we have, based on the runway we have to continue to drive both upgrades and adoption, I think we feel good about the sustainability of the growth we're delivering in that segment.
Yeah. We typically have customers push us for long-term contracts.
Okay.
They want those three-five-year contracts. They want to lock in. For us, we get that first-year uptick. Either it's an upgrade or we're bringing more products in, so we're moving customers, say, from Westlaw Advantage to CoCounsel Legal, which includes Practical Law Dynamic Tool Set and CoCounsel Essentials in addition to Westlaw Advantage. We get that first-year uptick, we also get the benefit of the year-over-year increases. We're layering more innovation, we're moving faster with our next steps, we can capture a higher year-over-year increase than we historically did.
Okay. Are there any questions from the audience? Just sticking with pricing. Definitely that's one where there's concerns in the market around pricing as AI evolves. I know you've had some discussions around value or usage-based pricing. I mean, where are you in the explorations at this point, and how do you think about pricing as the market evolves?
Yeah. First of all, our philosophy as a company is to price for value. I think there's a lot of products, particularly ones we've had for a long time, we've got a very good sense of the value it's providing. We have driven pricing commensurate with the growth of that over time. I think that right now, as Steve said, when you upgrade to a higher tier, you step up to a higher price point. With the newer, more innovative AI solutions. We are seeking higher out-year pricing. I think pricing is good. We've acknowledged, I think this may be more where you're going.
Like I think a lot of other companies in the software and information data services space, we are thinking through the potential to add a consumption-based component to pricing at some point in the future. I think it's logical on a couple of perspectives. The obvious one we're asked a lot about is we have a variable cost with usage around the LLM costs.
Yeah.
That's in particular the case with the agentic solutions.
Yeah.
I think there's some logic to do that. Now, there is no short-term margin imperative to do so, to be clear. The step-up in price we get and the faster out-year pricing that we've gotten historically, those price points were derived based on our understanding of and forecasts around usage and these kind of things. We are incredibly good at architecting our solutions to use models that are best suited on thinking about the combination of their accuracy, their speed to deliver an output, and also cost. I think we manage this cost quite well.
Yeah.
LLM costs also are not very big in the grand scheme of our cost structure. In light of that, and in light of different types of users likely using some of these capabilities with a different intensity based on a lawyer versus general counsel or other factors like that, we do think that moving to what will likely still be enterprise subscription-based fixed pricing, but maybe with some usage bands or tiers or something like that, is logical. I think the whole software and information services world is thinking about this. My guess is we won't do it in the next six months. I'd be surprised if we haven't begun to introduce that in the next two years across many of these newer AI-driven products.
Yeah. Is that something customers are thinking about here, or customers are still so early in the journey that they're not really focused on it?
I don't know, Gary.
We have not introduced this yet to customers and said we're going to do this. I think certainly we're in the process of socializing it, as are a lot of other companies.
Yeah, sure.
I think we will do it in a thoughtful way, I think in partnership or at least with feedback from our customers, if and when we make the decision to do that. I think our thought process at this point is the pricing will still largely be enterprise, enterprise-based.
Yeah, I guess I hesitated for a minute just because I wasn't sure what Gary had shared. Two, maybe I'll give another couple examples. For a number of customers who have CoCounsel Legal, they're actually contracting with us to actually configure workflows within CoCounsel Legal.
Okay.
We've done a number of those internally, but we actually just announced one externally. This is a Sterne Kessler partnership. Sterne Kessler partnered with us. We co-developed a patent claims evaluation analyzer, and it runs patent claims through a Section 101 analysis and tells a user whether it's viable or not bringing that forward. Customers might engage with us that way and incur additional cost, and in some cases, we actually structure a revenue share for them. There's additional opportunity there. Yeah, we're still in the early stages of conversation about the consumption pricing.
Okay. That sounds good. You touched on token costs in your answer earlier.
Just trying to front run your next one.
Yeah. You knew I was thinking about it. TRI on their last call talked about creating their own kind of large language model specific for legal.
Yeah.
Maybe you can talk a little bit about that initiative and more broadly, how you're thinking about working with the other AI model providers here.
Yeah. I'll have a couple of thoughts. We talked about Thomson, which is our large language model that we are developing, and we're doing that through a small acquisition we did in 2024 of a business called Safe Sign, and really it was a talented team of legal and computer science experts affiliated with University of Cambridge and Harvard University primarily. They were excited to join us to have access to our content sets. They had gone down the path before we acquired them of beginning to build a legal specific language model. We are really encouraged by what we're seeing from the model, in terms of accuracy, which on certain legal tasks has actually matched or surpassed some of the leading frontier models. It's also cheaper to run for us than using a third-party model.
In many skills or capabilities you use it for, it comes back with an answer significantly more quickly, which is important in the customer experience, reducing latency in these solutions. I think we'll triangulate around the concept of accuracy, speed, and cost, and I think we would hope, by the end of this year or as we move into next year for sure, that we will begin to build Thomson into certain capabilities or tasks that our AI solutions deliver. I think that will be helpful. We have talked a lot about being relatively model agnostic, meaning that we'll use the best models in our solutions, and we've built our AI architecture in a way that we can swap out models. We've done that a couple of times already in the last few years.
Currently, for our use cases, we find the Anthropic agentic capabilities to be the best solution. That's what's powering Westlaw Advantage, and that's also what will power the next generation version of CoCounsel that's currently in beta. Within that really is the orchestration layer that runs the agentic process. It could determine, hey, here's the 10 steps we'll do to deliver this answer to this customer. We can have other models actually actioning the various steps. I've gotten asked a lot lately, well, what if Anthropic raises the price? I think as I alluded to earlier, our engineering labs teams are constantly testing all these models and really good at optimizing our solutions for the concepts that I said before of accuracy, speed, and cost. I think if someone surpasses what Anthropic's doing, I think we would consider changing.
We obviously have a strong relationship with them and are really excited about what our products can do powered by their latest agentic models today.
Any questions from the audience? In the market around AI, Steve, how do you think from your point of view in the business, what do you think the risks and opportunities are around AI as you look out in that legal business over the next few years?
I think that there are just exceptional risks, and I think that's reflected in the financial results that I'm pleased we keep on delivering better and better results. If I think about the categories that my team plays in as an example, like traditional legal research, know-how, those are established categories. We continue to grow those categories. I think a major new opportunity for us is this emerging category of AI workflow solutions.
Yeah.
That's where I think the re-architecture of CoCounsel Legal, taking that from being more of a wrapper around a large language model and actually embedding Claude Agent SDK natively in our content so that it can reason in real-time and produce these near autonomous agents is a game changer. As we show customers that, in the early conversations where we showed them vision, as we now engage customers in the alpha and then the beta, the feedback is incredible because they haven't experienced something like that grounded in the data, trained in the data, with an agent truly moving in real-time. I think right now where the wrappers are at is they produce these workflow steps, and a lot of other vendors in the market have workflow builders, and they allow innovation teams to go in and create their own workflows for the team.
The challenge with those workflow builders is that over time, as the law changes, they get stale, and you need to go in and manually redo them. With our agents now coming out next generation CoCounsel.
They can change in real time. As the law changes, you see those agents change, the number of steps that they perform, the exact steps they perform. I think that's a tremendous opportunity for us. That said, on the risk side, I think I have an exceptional respect for the competitors, especially the startups that we have in the market. I think they're moving very fast. I think they're putting a lot of money into their user applications, the experts that they have supporting them, their sales teams, and I think they're going to continue to make a lot of noise in the market. I do think in places they're going to add real value, and I'm hearing good things from customers. The thing that I feel good about is that our tools continue to play side by side with them.
My customers will buy CoCounsel Legal, and then they'll buy maybe one or two of those other products. They're signing long-term agreements with me. They're signing maybe a one-year agreement with one or two of those others, and they're kind of experimenting with both and seeing where the value is. I also think there's an interesting movement in the market right now around the foundation models.
You have Anthropic expanding their Claude for legal sets of plugins. I was on stage last week with Mark Pike, who is the Associate General Counsel at Anthropic, and he is doing lots of very interesting things. They are not moving into fiduciary-grade AI.
Yeah.
Nor do they have plans to. He made that very clear. I think the risk comes from just competitors continually and new entrants every day doing interesting things. There's always going to be a place for us.
Yeah. When you ask risks, where I go to is we need to execute. We need to execute from a product perspective. We need to execute from an engineering perspective. We are moving and innovating faster than TR ever has.
Yeah.
The competition is doing that as well. I think if we innovate and bring our key assets and our content and our deep bench of experts to bear, I think we should be able to create solutions that are difficult for any of the competitors out there to compete with in this AI workflow space. We need to execute.
Yeah.
We also need to execute, Steve and our other go-to-market leaders need to execute well. Those of us in the center need to allocate capital well. I feel like a lot of our performance in the next few years is in our control.
Yeah.
We really believe that and are focused on that and are not arrogant about that. We know we need to continue to accelerate the pace with which we are innovating and delivering to win in these markets, but I think we feel really good about our ability to do so.
I have more questions than the five minutes I have time for, but are there any questions in the audience one last time? No. Let's move to the margin side. I think folks were a little surprised at the 100-basis-point margin improvement expectation. You guys already have very strong margins. Maybe you can talk through how to think about the productivity gains you're seeing and how to think about that 100 basis points.
In my mind, there's three categories. There's underlying operating leverage the business has, and I'll comment on that. There is some incremental productivity that we are delivering and excited about the potential to deliver more that we discussed on our fourth quarter call in the form of what we're calling Reimagining How We Work. It is focusing on more automation internally, a lot of that AI-driven. The third category is how much we choose to reinvest in high-return, exciting growth projects. The bottom line, before I give you slightly more color. Across the three, we're very confident delivering to the healthy margin expansion t hat we've guided for. Mike Eastwood, our recently retired CFO, used to say a few times on earnings calls, 60%-65% of our costs are relatively fixed in the short term. Think people, wages, compensation, think technology, subscriptions, that kind of thing.
Those are growing, I think, in the 4%-5% range over time. With inflation, they can move a little bit, but generally, I'd peg them in that range. If you said 5%, we're growing revenue this year, our guidance organically is 7.5%-8%. On those fixed costs, we are driving leverage. On the variable costs, I think it's a reasonable assumption to assume they grow roughly in line with revenue. There's a few LLM costs or cloud costs that might be growing faster. There's a number of them that are growing more slowly as well.
I think if you do the back-of-the-envelope arithmetic that suggests, there is very healthy underlying operating leverage in the business. That hasn't changed meaningfully in the last 12 months or anything. We did not provide specific targets for the Reimagine Our Work initiatives, but we have a very structured process internally where we've invested a lot in the AI tools, in training, and we've, think, fostered a culture of experimentation throughout the company where our employees are focused on this. Then last year, we layered in a top-down structure to manage and sort of drive this transformation, our productivity initiatives. There's meaningful savings there. The upside to those two to us is it leaves us room to reinvest pretty meaningfully, and we're doing so in a lot of categories, and we'll continue to while having the ability to deliver bottom-line margin expansion as well.
Okay. Let's end on capital allocation. TRI recently announced an additional $600 million in share repurchases, a $605 million in return of capital. How are you thinking about capital allocation at this point in terms of the various buckets?
Yeah. We've talked over the years about having a balanced capital allocation approach. The first non-negotiable is reinvesting in our business, right? We're blessed with a strong balance sheet, strong cash flow, and so that will always happen. We are committed to growing the dividend, as you've seen. Each of the last five years, we've grown the dividend by 10% a year, and I think this year was our 33rd or 34th consecutive year of dividend growth. Strategic M&A remains a real priority, and I think over time will continue to be a big focus. The hurdle rates there are high, not just sort of strategically and financially, but we also focus a lot on cultural fit. We don't want to bring in a lot of tech debt, so we're typically looking for modern technology.
I'd say there's a fifth sort of hurdle, as we think about M&A today, which is AI disruption risk.
Yeah.
Yeah, I have a hard enough time with all of you and all of you dealing with the narratives that we're facing. We don't want to open a new front in that battle by acquiring something.
Yeah
We weren't really confident had a strong ability to win from AI. M&A remains a real focus, and I think we'd like to do M&A. We also have a history of repurchasing our shares, and today, that is particularly attractive given where the share price are, where interest rates is. When we run the returns analysis we do for M&A, for buybacks, and for large internal investments. Buybacks look attractive. I think that was the primary reason we did the $1 billion to return a capital buyback early this year. While ultimately, these are board-level decisions, I'm confident our leadership team will continue to discuss capital allocation with the board, and that we will continue to have this balanced approach of delivering against all of those priorities.
Can I just say one thing in closing?
Absolutely.
The real reason why I came here and the message I want to leave is, I've been with the company for 15 years this week. I've never seen this level of excitement in the customer base, in our sales teams, in all of our back-office professionals as I do right now. My group has never grown faster in its entire history. I can't remember exactly what we disclosed on the last earnings call. I think we said double-digit growth.
Yep
Which we are seeing. I think that's a reflection of just the excitement, the partnership with our customers, how excited they are about what we're seeing in the market right now and the opportunity ahead.
Perfect. All right. Great place to end. Thank you very much, gentlemen.
Thank you.
Thank you.