Thomson Reuters Corporation ($TRI)

Thank you for being here. For those of you who don't know me, my name is Manav Patnaik. I cover business and information services with Barclays. We're pleased to kick off the afternoon session here with the team from Thomson Reuters. As you can see on the screen, we've got Steve Hasker, who's the CEO, but we also have Gary Bisbee, who's Head of IR; and Kirsty Roth, who is the Chief Operations and Technology Officer joining us here. So we're in battleground mode here, so bear with us.

Steve, maybe just to start at a high level, given all the debate, the noise. Maybe it's worth just taking a step back. And from your perspective, how do you describe Thomson Reuters and its positioning in its moat today to a new investor.

Yes. I mean we've been saying for a number of years now that we have a nice tailwind behind us, which is the growing complexity associated with compliance. So that's legal compliance, tax and accounting, audit-related compliance. And we're one of the few companies that have the content-driven technology to enable companies and their advisers to meet or exceed the needs of those compliance requirements without adding a lot of headcount. So that's the first tailwind, and that's helped drive us from when Kirsty and I joined the company 5, 6 years ago, we were growing our Big 3 at low single digits to now we're in high singles, and we've got line of sight to, we think, double-digit growth in the years to come for the Big 3 franchises. I think the second tailwind for us is AI. We've been in the advanced machine learning space for 30 years. We put a functioning search algorithm on the front of the Westlaw data set in the early 1990s before Google was commercialized. So it's not -- when ChatGPT passed the bar exam, it wasn't sort of entirely new territory for us, and it's certainly something that Kirsty and her teams have worked on before that, and very intensively since. And essentially, what agentic AI enables us to do is play a much larger role in the success of our customers. So in other words, for a tax and accounting professional instead of merely being the tax calculation engine, we can now provide an end-to-end agentic solution which covers all aspects of the tax return process. And then for a litigator, instead of being the primary litigation research data set, we're now in the drafting business. So whether that's contract, a motion to dismiss, a motion to compel, a rebuttal and so forth. And so the reason we are confident in this AI environment, despite all the noise is our positioning. So we're the only company, we believe, that can provide fiduciary-grade AI. And this is a term we've coined in the last couple of months, particularly with the sort of emergence of the horizontal players like Claude and Gemini and so forth. Fiduciary grade is basically serving those professions where they cannot afford to be wrong. They cannot afford to hallucinate. And for those of you who follow the legal profession, you will have seen that several of the global large law firms have stumbled very badly in the last couple of months in the context of big transactions, big litigations whereby they have used a generic chatbot, produced the wrong results and then had the opposing counsel, in most cases, supported by Thomson Reuters AI solutions, point out the holes. In other words, pull their pants down in front of a judge or in front of their clients. And so the reason we are advantaged in this environment is, firstly, we have the content. So 85% of our content is not publicly available. It's been created and curated by our experts. Secondly, we have the experts to help train our agents. And Kirsty's team have done a wonderful job of taking all of that expertise out of the heads of 2,700 deep practice experts and ensuring that our agents, including CoCounsel Next, which was launched in beta a couple of weeks ago, behave like a highly tenured expert. Thirdly, we've got data privacy and protection protocols that no 1 else has. And what I mean by that is we make a promise to our customers, which is their input will not become part of our AI output. And that is quite unique in Silicon Valley, right? Everyone else is sort of putting an agent out there, putting a chatbot out there and learning from the queries, we are not. And the reason we're not we don't have to is because we have our own experts in-house, and we have our own content. And then last but not least, we've got exhaustive customer support. And the reason I point that out, it sounds quite mundane. But if you use Westlaw Advantage, our AI, agentic AI product or you use CoCounsel, and you get a result that you don't understand, you can call a highly tenured attorney, live 24/7 and he or she will have access to the query that you've run and also has a deep understanding of that particular practice area. And you can ask them, "Well, why is it producing this result? I mean, in front of a judge tomorrow, I cannot afford to be wrong, I don't understand the result that's coming up." And you can track and trace it all the way back to its source and the same in the tax accounting the same in the order professions. And you've got to ask yourself, are the -- some of the start-ups or the frontier models themselves going to have that level of personal support. I mean they're not, right? And so we think that, that positions us to coexist with the frontier models. But to be the only player who can truly serve this fiduciary-grade AI. And that's why we're bullish and we're particularly bullish on the back of yesterday's results where our Legal Professionals business ex the government component grew by 11% organically. That's the highest we've seen. And we're confident that we can sort of drive this AI road map through our offerings and see accelerating organic growth.

Got it. Kirsty, maybe from your standpoint as well, you were a Chief Transformation Officer when you guys first created what's now just legal and tax accounting. What is the biggest change or your competitive advantages from your standpoint from your tech role as well?

Yes. I mean if you think about it, right, lately. -- all roles are impacted by AI. And certainly, whilst as Steve described, we're putting a huge amount of effort that we put in the products, we're also changing the way we operate as a company. And I just think that has huge implications, right? Like the speed with which we can -- we're launching features faster than we ever have done. We are putting out new products in market, we're really able, as Steve described, therefore, to help our customers move from if you take tax, just a world where they're returning -- providing returns to getting into advisory services, right, because you crunch the amount of time they're spending on a return. And so the up leveling we've seen in early days, particularly in the engineering and the support teams for how much work they can get through that means we can move more quickly, I think, is sort of super exciting. And then we'll see that sort of through every function, Manav. And I think that also gives us a real advantage. And we started that kind of work about 2.5 years ago. And I talked to a lot of peers who are COOs of -- I call as all legacy companies, right? All those kind of companies, and I know we're in a really good spot as to what we're doing. And I think that's really important. I don't think you can be in this world and be standing still and just taking advantage, you've actually got to reinvent the whole way you operate right now so that you're moving more quickly and serving your customers even better. And that's exciting because we're seeing real results.

Got it. And talking about real results, Gary, maybe firstly, thank you for reporting and getting on a plane right way to come here. But maybe just help us with last quarter -- last night's results, a quick summary. I mean it wasn't perfect. It was one of the stronger quarters you guys have reported. So if you can just quick highlights there?

Yes, yes, for sure. So we reported 8% organic revenue growth across the total company. We had -- that's up from 7% that we reported throughout 2025, and we had guided for the quarter for 7%. The outperformance was driven by our Legal business and our Corporates business. In Legal, we -- as Steve referenced, we grew the law firm customers, so that excludes the government customers by 11%. That was up from 9% last quarter. And the acceleration there is driven by the tremendous success we've had with the launch of Westlaw Advantage, the new top tier that we launched last August and continued strong growth of CoCounsel, the AI system product. And so I think we feel really good about the result. We also as Steve said, discussed fiduciary-grade AI and shared for the first time some incremental customer usage data that I think supports the story we're telling and we're seeing, which is that we're seeing strong customer adoption and usage of these products that you see flowing through to better revenue growth. And so for the full year, we reiterated our guidance across revenue and profits and feel really good about the start to the year we've had and the outlook for the business, particularly on the back of that strengthening in the Legal business.

Steve, maybe let's start talk about some of this noise that you're alluding to. So maybe let's start with the Claude releases. Last night, there was something around the financial plug-ins in February, specific to the Legal plug-ins. There was obviously a crazy reaction in February. Harsh reaction yesterday, but just -- but then you were on stage with Claude as well. So what's really going on? What's the partnership...

Yes. So the way we've built our AI offerings were model agnostic. So we can and we do chop and change between whatever is the best frontier model at any particular time. So for CoCounsel Legal, that's Claude Sonnet 4.6, now Sonnet 4.7, and it may change as things go on. For our tax accounting products, that has been the latest ChatGPT models. We also have built our own model. It's called [ Thomson ], and that's going to go into sort of full release as we go through this year. And the reason I point that out is it's a large language model built specific for legal. It's entirely owned and operated by us, built by a number of former Google deep mind scientists who are on staff. And starting to outperform the very best of the frontier models in terms of speed and accuracy for certain legal tasks. So that gives us a level of optionality as we go forward. So we have a relationship with Anthropic. They're a big vendor to us. We've done a lot of co-development with them. And I think there are opportunities for us to expand that set of arrangements in various ways. I think though there's a sort of an assumption from investors that these frontier models are going to sort of eat all before them, right? They're going machine learn their way into the content sets, they're going to be able to perform with sort of great levels of accuracy. We are very [ family ] of the view based on everything we see, the customer interactions we have and the accuracy of our products that this fiduciary grade is a spot that a general purpose model can't and won't meet that standard. If you're a lawyer, you cannot afford to be wrong. If you're a tax accounting professional, you cannot afford to be wrong. The fines, the risk of losing your license, the loss of client relationships. All of those things mean that you need to have CoCounsel in addition to a Cowork plug-in or a Copilot plug-in or a Gemini plug-in or whatever it might be. And we're pretty confident that, that's how the market will evolve. But I think at the moment, investors have sort of decided, well, if you're a native AI company in one of these frontier models, you're a winner. And if you've been around for a while, you're a loser. And of course, and this is why Kirsty's role is so important. The winners will be those who execute. I think our product road map is fairly clear, at least to us and to our customers. And the rate of innovation is accelerating at Thomson Reuters. And for me, the only thing we're watching is just can we go faster and can we produce better and better and more innovative products as we go. If we do that, we'll emerge not only as the unequivocal winner as we have been over time in legal research in the tax calculation engine space, we'll also move into this broader workflow space, and that's a new TAM for us. And that's why we're confident we can see this accelerating growth rates.

Got it. And to your point and move faster, I mean, some of this might just be perception, cars, you can jump in as well. But when the fintech competitors or the legal tech competitors are talking, people think that these are the smartest guys out there. But I think a lot of your progress is buried on...

Yes. Can I give you an example of when I joined the company 6 years ago, we used to put out a new version of Westlaw sort of a couple of times a decade. It was a meaningful upgrade at each time and price premium associated, [ et cetera ]. Now we put out a new version every 4 months. And we started the work on Westlaw deep research, which is a fully agentic product, at the same time that ChatGPT's deep research product came out. And we're able to outrun their deep research product in terms of its utility for high-end legal practice. And so this is something where we've got to earn the right every single day in terms of our rate of innovation and the success of those innovations. But I would -- for the professions that we serve, so tax, accounting, legal, audit, risk, fraud and compliance, I'd stand up our rate of innovation against anyone.

Got it. And MCP is a big buzzword right now. Can you just clarify like your partnership with these LLM and the monetization model and what's actually being shared?

Yes. So at the moment, they're fairly vanilla vendor relationships. So we haven't gone into sort of complicated rev shares or any of those kinds of things because it really is a case of picking the -- doing comprehensive evals, picking the best model, utilizing that at that particular time. I think there will be options going forward. the asks we get very regularly from all manner of players is they want to get access to our content and to our experts. And we see no reason to share that IP with anybody at this point in time.

Fair enough. And then maybe shifting to, I guess, you guys -- there's the content and research side of the market, and there's the workflow side of the market where it's more of the debate. So I know, Steve, Kirsty, both of you talk to clients a lot. So can you just help us peel back the onion behind the scenes? What are the lawyers, your counterparts telling you, but Counsel versus Harvey, Legora, all these other alternatives out there, what's really going on?

Well, I'd say the first thing is at the moment, the tools are 18 to 24 months ahead of the change management within the firms. So I think there's this sort of narrative that these tools are there, it's all said and done the picker profession that's being transformed. I think within software coding, that is a real -- a simplistic but realistic picture. When it comes to legal professional, the tax and accounting profession, it is very early days. And so what we find amongst our customers is we have a number of customers who are Thomson Reuters only, and then we'll have a number of customers who are trialing on a paid basis, Harvey or Legora or some other tool and ours as well. And as I said, we're confident, particularly on the back of the Westlaw -- or the CoCounsel Next beta test that's now going on. that will emerge amongst that pack for the reasons I described. But at the moment, there's a lot of experimentation and there's -- and the firms are just starting the underlying change management, which basically means rewiring the way in which the work product is produced and the interactions between a senior lawyer and a partner.

So maybe it's the experimentation that's a disconnect because I think the view is that Harvey is growing whatever, Legora, so that means you've lost that opportunity. But it sounds like these are shorter-term contracts?

Yes. I mean, look, this is a market we weren't in, right? So I think it's interesting to see. And to Steve's point, I mean, I was actually with a coal group of the sort of COOs of many of the biggest law firms yesterday. And these are companies that have never really been through a huge amount of change. They're not greater change anyway, right, even compared to Corporates that are often not [ greater ] change. So I think it's just fascinating to see. They've got all these tools they've bought to sort of feel like they're trying stuff out and feel like they're taking a box, but they don't quite know what they want them to do. I think what they do know and where we get a lot of feedback is people don't really want to buy 5 things. So they might buy a Claude or a Copilot for the generic things that are sort of doing office admin, et cetera, [ they're really ] like one tool. And that's partly why we've moved to sort of packaging, CoCounsel's gives you access to Westlaw and Practical Law and everything and sort of that -- an old [ speaker ], sort of single pane of glass that does everything for you, as I think that's going to play very much in our favor. But honestly, the change management right now is it super early days, I would say, in these firms and a lot of -- where I'm spending time with them is to help them figure out how you do that. because it's not just about which tool you want to be using. They've got to think about their data and their integration and how that works. And there's a lot of anxiety to Steve's point earlier on, they can't afford to be wrong. Actually, one of the interesting discussions yesterday was, obviously, they're getting GCs or the general counsels that are obviously using many of these tools and then sending them wrong inputs to do their work. And that's not great either. So a lot of concern. I think Steve mentioned many of our winning point earlier, but one get involved with a lot is the security and privacy. Our customers can't afford for that data to end up in the wrong place, and they want to work with partners that can do that. So really think about how they put their whole sort of integrated suite of products together in a way that's effective for them. So a lot for them to do. And I don't think that -- I think that's going to be a multiyear journey.

If I could just add on. In our simplistic way of thinking about it, we have our legacy markets of content-enabled technology. So Westlaw in legal research, practical law, content around transactional areas, [ a lot ]. The competitive dynamics there haven't changed a lot. It's a duopoly for the most part. LexisNexis is a strong #2 in both categories. We're both innovating with AI. We think we're ahead there, but we're seeing that drive growth as we drive upgrades to these new AI-enabled offerings. Then you've got another market, which we think is a significant incremental new TAM white space for us. That's AI workflow solutions. There are more competitors there. We're one of the leaders in that. We've had strong growth in that market. The question for us is how much of that incremental TAM do we win? And then how impactful is it on our growth rates. And we've been a leader to date, we think with the next-gen version of CoCounsel, which more powerfully uses our content for the AI to do more. That's our play as we execute to be a clear winner in that market going forward. But even if we're one of a handful of players playing there. It's a significant incremental growth opportunity for us.

That makes sense. And you guys obviously gave us some user stats on CoCounsel recently. But obviously, some of the legal tech competitors give us like an ARR number and growth rate. Anecdotally, any way to compare CoCounsel...

We haven't sort of come out with a precise number as some of those companies have. House is in the same zip code. So we're notwithstanding that the CoCounsel Next that Gary just mentioned, is really the first time that we've harnessed the full power of our content sets and our expertise to train that agent, that CoCounsel agent. Notwithstanding that, we're at 1 million users. We're in an ARR that's similar to some of those players. So we've made a good start and now watch us accelerate.

I mean I think 2 dynamics to add. One, the main sort of fireside lobbing I got yesterday was people are hearing the beat is really good and they want to get on the beta. And obviously, we're still in beta phase until the summer. And the other thing is we're now solving about 3x more problems. So for many of our clients, they'll tell us we love CoCounsel, but it doesn't do this and that and the other, and we'd like it to do more. The folks who are in the beta are now immediately going back and retesting those things, they would hope it would do and it didn't do them, and it does them all. And so that's literally getting a sort of 3x uplift on the things that it can do. So we expect that to accelerate very quickly. I think in my day job, I'm going to have a scalability challenge us to make sure we can go fast enough to keep up with demand. That's certainly what keeps me up at night at the moment.

Got it. So outside of the workflow, the other debate, like you said, I mean, you're in a basket and nobody listens to the hallucination stories that come out every day, which should help you. But the other debate, obviously, is the data, right? Like you said 85% of it is close to proprietary, fiduciary grade, can you just double-click on that? Like talk a little bit more about where the raw data comes? How do you get to 85%?

Gary, do you want to take that one? He and I have debated this over time and I know it's near and dear to your heart.

Yes. So the first thing I'd say is we went in quite a bit of depth on our third quarter conference call in November. And so there are a couple of, I thought, reasonably compelling slides and 5 minutes of Steve's talk time on this. So I'd point people to that to understand this. But we think there's a significant moat around Westlaw and our other content assets. And I'd just make a couple of points. One, a lot of the data is just not publicly available. So when a court issues a ruling, the judge will put out a summary judgment effectively a summary of the outcome and what happened in the case. We get lots of other content briefs and other things from the courts and from other sources. There's also a bunch of it, like the docket information that's behind paywalls. So the models can -- a competitor could buy those but billions of pages. At $0.10 a page, it's a significant cost. Then all the data we get is completely unstructured. We get it from more than 3,500 sources. We process more than 300 million documents a year into Westlaw. And there is a significant process, people, technology to standardize and be able to use that data how you organize it is really key. We've got a ton of software tools that we've developed that are part of this. Things like the key number system, it's our taxonomy, or indexing system, 140,000 legal matter deep updated twice a year. Everything in Westlaw, from summaries to head notes to citations are all organized against that. That is a difficult thing to do, and it is why we can be both comprehensive and accurate because it's all organized that way. Things like KeyCite. It checks or citations. It also tells you when a case has been overruled and you need to know that, right? It's not just about, hey, a model could go find something online about some case that's relevant. Is that still valid law? And a lot of those tools we have to wind this down and not talk too long, are really key to the final step of anyone that does legal research, which is verifying the outcomes. That's incredibly key with an output where they can make it up. And so we have the best tools around verification, and that is something that our customers, I think, really need. There was -- Steve alluded to this, there was a very public, a couple of weeks ago, incident, we're a large law firm turned in a brief that had some hallucinations in it. And the other side that found the hallucinations and showed them to the judge use Westlaw Advantage. But the other teams brief into a tool we have in there called Litigation Document Analyzer. It said, here's the 4 mistakes, they turned it in. This is a great example of this fiduciary-grade AI, really hard to do. We've spent a ton of time and effort, process technology, people to create not only the content but the tools that deliver this, and we think that would be incredibly difficult to replicate.

The other debate within the data we get is like relevance of the size of the firm and if they're doing transaction law versus case flow? Like how much of them really need this data? How do you approach that?

Well, law firms, I mean, they're always looking for new clients and new growth opportunities, and they are prepared, therefore, to extend into new practice areas, where they don't have necessarily a corpus of information in their own document management system or in the minds of their partners and associates. So one of the things that we bring in both Westlaw on the litigation side, Practical Law on the transaction side is a comprehensive across jurisdictions, across practice areas, robust, up to date. And so even if a firm is really deep in a particular area, chances are they're going to want to go elsewhere and they want that source of truth. So that's -- and there's a set of investments that we just continue to make.

Got it. And if we kind of moved a little bit into the Corporates segment, sticking to the legal side, like how do you think of the opportunity there with the...

Yes. So it's a great question because the sort of -- the heritage of Thomson Reuters Legal Professionals or Legal business has been Westlaw, which is litigation, and that's the law firms predominantly. I mean, they're obviously general counsels [ are in ] industries where there's a lot of litigation. But by and large, the litigation practice tends to focus on the outside counsel. And so that's been the sort of heritage of the company. The opportunity with general counsels for us is we've been chipping away at it for a few years. but it is to a very large degree, white space. And what we're starting to see is general counsels who perhaps work with a CTO who's saying, okay, to Kirsty's earlier point, let's use Copilot or let's use Gemini for various sort of aspects of knowledge work. But the general counsel and for that matter, the Head of Tax or Head of Internal Audit, needs a specialist tool that ensures they're accurate. And they come to TR because we've already got a set of relationships with their corporation, particularly through the ONESOURCE suite. And so we think there's a pretty big opportunity for us to grow and become the leading provider of that workflow solution set to the general counsel. And in doing so, help them drive significant efficiency and effectiveness through both their in-house and external spend.

Got it. That's helpful. Before we move to Tax & Accounting, just on the Corporates side, towards the end of last year, you talked about realignment in the sales and creating a little bit of a timing hiccup...

Yes. [indiscernible] basically, what we did was we sort of moved more of our activity to a global accounts model. And in doing so, we upset sort of territory and quotas for a bunch of salespeople. We're now 12 months sort of through that, and we're starting to see that really settled down. So we like the look of that. We've had some big wins among some of the marquee sort of Fortune 10 and Fortune 50, both in the general counsel's office and particularly in the indirect and the e-invoicing space. So I'm optimistic about the sort of rest of this year into next.

Got it. A lot of those basket traders are not giving Legal a shot. They don't give Tax & Accounting a short either. So can you just help us with what the moats there are and why that's growing so nicely for...

Yes. I'll start, and I'm sure Gary and Kirsty will add. But essentially, we are in duopolies in terms of the tax calculation engine space. So as it pertains to serving tax accounting professionals, so tax preparers, we and one other player have sort of pretty equal shares in that space. And then when you go across to the -- serving the Head of Tax and the Big 4 tax and accounting firms, we're the largest provider of direct and indirect. So direct being income tax and indirect being sales, excise, VAT and so forth. We made an acquisition a couple of years ago of the leading e-invoicing supplier. So as e-invoicing mandates are rolled out by tax authorities across the world, we're seeing tremendous growth from that asset in Pagero. And so this is a space where the sort of moat to use your term is around the fact that these tools are extremely accurate. I think we spent a lot of time and effort making sure that all of the [ imported ] information is perfectly accurate. They are very cheap to run and they're very efficient. So there's not a lot of incentive for a Head of Tax between tax filing systems to go tear that tax calculation engine and try something else, right? Because we've got the last 7 or more years of data in that. So you've got that sort of time series of tax return data. You've got a team that are used to using the system, and you've got all the interconnects to the IRS and the other tax authorities. So it really is a bit like concrete. And the opportunity for us is to add the agents in terms of sort of on the shoulders of that, so that the document ingestion is fully automated, all the way through to the e-invoicing and the [ cues ] for advisory services. And so that's the ready-to review proposition that we put in the marketplace earlier this year and the ready to advise proposition we put in the marketplace last year in a series of CoCounsel agents that are around it. So we think that's a deep and broad moat and it gives us the ability to drive fundamental efficiencies throughout the tax preparer space by adding agents and automating some of the things that take a long time and candidly, a sort of treasury in terms of the work.

Got it. And it's also -- I'd add one thing to that, which is this concept of automation is key for this industry because the number of people coming out of university with accounting degrees is down. The number of people sitting for the CPA exam down by like 1/3 in the last 15 years. And so this industry needs technology to do more. Number of returns goes up every year, a number of audits goes up, the complexity of both is rising, and yet headcount is in the best case, stable. And so in some ways, in some ways, the propensity to adopt new technology, including our AI offers, one could argue is actually a better in that industry than legal.

Got it. investors are often surprised when I tell them that the pricing and tax accounting is firmly in the mid-single digits and more than legal. But then the counter is because of AI and solutions, it's going to be price deflationary. So how do you guys think of pricing in T&A and across?

Well, as I said, I think these tax calculation engines represent a very significant moat. There's neither the incentive at the customer side to experiment with a new mousetrap nor is it an easy thing to do. And it's not something that AI sort of cracked the code on. And so I think we -- our rate of innovation and adding the agentic capabilities or to outpace sort of any start-ups that look to get into that space. So we think that there's both -- we can continue the sort of pricing component, but also there's a quantity component. That if we play a larger role in that [ broaden to end ] tax prep process that we're saving significant amounts of money for our customers. And therefore, with value-based pricing, we'll extract some of that.

Think about sort of the macro of what's going on, right, whether you're a CPA or you're a corporate, basically, now everyone is looking at it in roles like mine and saying, okay? I used to spend 100% or 95% of my cost on people and 5% on software. Now I'm going to shift it. And so actually, surprisingly, very little or very rarely does this come down to price. People are just making a different choice. They're not going to hire as many people, they are just going to spend more on software. And so I think that gives us a really nice sort of new market to sort of step into. So it's quite a different dynamic to probably what we even saw 12 months ago. So that's how people think about it.

And Kirsty, actually this time last year, I remember you're saying you were more excited about some of the AI stuff going on in Tax & Accounting than in Legal. What's the latest?

Still hugely excited. So I think as Steve described, I think we've acquired all of the pieces that mean you can go from sort of -- we all pay taxes, right? So we all know that lovely joy of how long it takes you to get everything together right the way through to filing. But when you talk to CPAs, that's, a, not where they have the most fun, it's also not where they make the most money. So they want that process to be as effective and accurate as possible. but they want to get into the advisory side of the business. And that's why we launched the product, Ready to Advise. And so they want to be able to come back to you and say, "Look, Manav, you're spending too much on this or that and you could do these other things." And there's an awful lot more revenue for them to pick up with those products. So it's, again, a whole sort of brand-new space for us where I think there's a ton of opportunity. So that's what I get excited about.

Got it. Maybe in the last 5 minutes, if you can touch on capital allocation. I mean, we know the dividend, we know the buyback philosophy. We can touch on that later to you have time. But more on the M&A side, Steve, I mean, you have generate a lot of cash flow. You have a lot of balance sheet flexibility and you talked about $9 billion available capital. How has your thought process around M&A change maybe a year or so ago, you would consider buying a software company? I'm sure you're not doing that today. So just what should we think about the priorities there?

Yes. I mean we've -- as you said, we've got about $9 billion in dry powder. So that's the sort of cash on hand plus the additional leverage we can put on plus the natural free cash flow generation of the business, which will hit $2.1 billion or so this year. And that affords us, I think, an advantaged position in this environment where there's a dislocation in terms of the valuation of a bunch of businesses. And I think a lot of uncertainty as to how this sort of AI environment will play out. We have, if you look over the last 6 years, we've been, I think, very disciplined. So we haven't done any sort of any outsized deals. We -- certainly, I don't think [ we've not have done any ] head scratches. We've basically stuck to the following criteria, which is let's focus on the big 3, so that's Legal, Tax, Corporates. Let's look at products that are additive to the customer experience within that group, products that are pristine in terms of the tech stack. So we don't want to require tech debt. And that typically means that they are AI-native products. Products that are, of course, accretive in terms of their financials to our shareholders, not just the outgoing shareholders. And then last but not least, a culture that is additive to ours. And that's not to say they need to look and feel like TR people. But we want to make sure that when the businesses come together, there will be a real focus on better serving our customers from the combined teams. And so if you look across the portfolio, Pagero, SafeSend, SurePrep, Additive, Materia, Noetica. These have been relatively small deals that have sort of pretty seamlessly plugged in. The playbook tends to be to buy those products and then push them through our distribution. I would expect, Manav, that we will continue to do those. We've got a keen eye on valuations, just given the dislocation I described. We are open to doing something bigger. Again, you're not going to see any head scratches from us. But we're open to doing something bigger. But at the moment, I think there's still a difference between the bid and the ask, particularly of those assets that are owned by private equity. To the extent that sort of the world has changed in the public markets, there's still private equity firms, the multiple is still up here. And so you'll probably see us continue to do the sort of singles and doubles and be patient in looking and opportunistic in terms of anything bigger. But Gary, Kirsty, anything to add?

No, I would agree with that. We'd like to do something somewhat bigger, but we don't have to because we have such strong underlying momentum, so we can be patient and I think we think that will work well. And to your framing, I would say one thing. We just had an investment committee last week to discuss a bunch of things. The first question we ask now, what is the AI disruption risk? Because we don't want to take on more and we need to make sure we understand that it has to be as durable as we see our existing franchises.

Okay. Maybe just to wrap it up then, $9 billion is a lot. You don't need all that for singles and doubles. And so we always get the question, why don't they just lever up and buy back shares, but there is some consideration...

Yes, we're open minded about it. I mean we recently completed a $605 [ million ] return of capital. which is tax deferred for the Canadian shareholders. And we're 1/3 of the way through a $600 million NCIB. And so when the Board meets again in a month or so's time, we'll sort of relook at that. We could certainly do something bigger. But again, it's weighing up being opportunistic and focused in terms of the M&A versus returning to shareholders. Certainly, at these share price levels, it's highly accretive.

Fair enough. All right. Let's see leave it there then. Thank you, Steve, Gary and Kirsty, for your time. Thank you, everybody.