Thomson Reuters Corp (TRI) 2025 Q4 法說會逐字稿

Good day, everyone, and welcome to the Thomson Reuters' fourth-quarter earnings call. Today's conference is being recorded.

At this time, I'd like to turn the call over to Gary Bisbee, Head of Investor Relations. Please go ahead.

Thank you, Melinda. Good morning, and thank you, everyone, for joining us today for our fourth quarter 2025 earnings call. I'm joined today by our CEO, Steve Hasker; and our CFO, Mike Eastwood, each of whom will discuss our results and take your questions following their remarks. (Event Instructions)

Throughout today's presentation, when we compare performance period-on-period, we discuss revenue growth before currency -- revenue growth rates before currency as well as on an organic basis. We believe this provides the best basis to measure the underlying performance of the business.

Today's presentation contains forward-looking statements and non-IFRS and other supplementary financial measures, which are discussed on this special note slide. Actual results may differ materially due to a number of risks and uncertainties discussed in reports and filings that we provide to regulatory agencies. You may access these documents on our website or by contacting our Investor Relations department. Before I turn it over to Steve, let me provide two quick updates. First, we have tweaked the naming for two of our segments to better represent their current focus and operations. Tax & Accounting Professionals is now known as Tax, Audit & Accounting Professionals, and Reuters News is now Reuters. The changes are name only with no impact on the composition or measurement of their results.

Second, please note that starting with our first quarter 2026 results, we will be making minor changes to where certain product revenue sits within the Big 3, reflecting how certain customers are managed. There is no impact on our consolidated results. For your convenience, we have posted to the Investor Relations section of our website a schedule that reflects our revised full year '23, '24 and '25 results and 2025 quarterly results in the manner we will begin reporting next quarter.

Let me now turn it over to Steve Hasker.

Thank you, Gary, and thanks to all of you for joining us today. Let me begin by commenting on the AI competition concerns that have led to recent share price volatility. We have growing confidence in the value of our content and our expertise for delivering professional-grade AI solutions. Our second quarter call in August, we discussed why Thomson Reuters is well positioned to win with Agentic AI. We bring comprehensive proprietary content, deep domain expertise and leading workflow software, which we combined with advanced reasoning models to complete complex multistep tasks.

Our unique and high-quality content grounds the AI outputs and our deep subject matter expertise trains and fine tunes the AI. Professional-grade results cannot be delivered without this content and expertise. On our third quarter call in November, we discussed the durable differentiation we believe exists with Westlaw. Its unmatched breadth of proprietary, editorially enhanced legal content, together with sophisticated tools to verify the AI deliver the comprehensive, accurate and up-to-date output needed for high stakes litigation. General purpose models cannot meet this standard.

Since those discussions, our optimism around Agentic opportunity has strengthened. The launch of Westlaw Advantage has gone extremely well with early sales and customer feedback indicating we have created a new standard in legal research capabilities. And more importantly, we have proven that we can leverage Agentic deep research capabilities to unleash our content and our expertise to clearly differentiate our AI solutions.

We are working to bring these advanced Agentic capabilities in the full power of Westlaw and Practical Law to CoCounsel legal by midyear and to our other legal tax and risk offerings in the future. We see legal AI workflows as a significant white space opportunity for TR and one that is largely incremental to our traditional research and know-how stronghold. We remain the clear leader in research and content. And we believe that we are a leader in the AI workflow market today.

Looking forward, our strategy is clear. We will continue to aggressively innovate with a focus on leveraging our proprietary content and deep domain expertise along with the latest AI capabilities to deliver highly specialized Agentic workflows that we believe will be difficult for those without the same content or expertise to replicate. I'll now turn to the fourth quarter.

2025 was a year of continued progress at Thomson Reuters. So let me start by reviewing some of our key accomplishments. First, we delivered another year of good financial results meeting our key targets. Full year organic revenue grew 7%, driven by 9% growth for the Big 3. Our adjusted EBITDA margin expanded by 100 basis points to 39.2%, meeting our outlook, and we delivered $1.95 billion in free cash flow, slightly ahead of expectations.

2025 saw continued acceleration in our pace of innovation with several foundational product launches, including Westlaw Advantage, the CoCounsel Legal unified solution and CoCounsel for Tax & Audit to name a few. Initial customer feedback and sales activity across these offerings has been encouraging, and we're excited about our road maps for 2026 as we work to bring deep research and Agentic capabilities more deeply into our portfolio.

Overall, commercial momentum across our AI-enabled offerings continues to build, highlighted by several large CoCounsel wins, including Microsoft, an important validation of our strategy and the market shift towards trusted domain-specific AI. We're also excited about initiatives we are driving to leverage AI internally to reimagine how we work, fostering a tech plus talent mindset to drive productivity and speed.

As I will discuss in a moment, we have delivered several early successes, including in software engineering, customer service, content operations and our General Counsel's office. This progress provides confidence in the long-term opportunity to become a more productive and nimble organization. Our capital capacity and liquidity remain a key asset we are focused on deploying to create shareholder value. In 2025, we invested $850 million into M&A including four strategic bolt-on acquisitions, bolstering key franchises and talent. We executed a $1 billion share repurchase program last fall. And this morning, we have announced another 10% increase in our annual common stock dividend.

We remain committed to a balanced capital allocation approach, and we continue to assess a number of inorganic opportunities. Our estimated $11 billion of capital capacity through 2028, we are positioned to be both aggressive and opportunistic.

Looking forward, our conviction around the medium-term growth potential for Thomson Reuters remain strong. As Mike will discuss in more detail, we are reaffirming our 2026 outlook for organic revenue growth of 7.5% to 8% including approximately 9.5% for the Big 3. We expect healthy margin expansion and strong free cash flow even as we continue to invest in innovating, serving our customers.

Now to the results for the quarter. Fourth quarter organic revenues grew 7%. Organic recurring and transactional revenue grew 9% and 8%, respectively, while Print revenues declined 6% in line with expectations. Adjusted EBITDA increased 8% to $777 million reflecting a 110 basis point margin increase to 38.7% due to healthy operating leverage and good cost discipline.

Turning to the fourth quarter results by segment. The Big 3 segments delivered 9% revenue growth. Legal organic revenue again grew 9% despite softer government growth, continued momentum from Westlaw and CoCounsel were the key drivers. Corporates organic revenue grew 9%, driven by offerings in our legal, tax and risk portfolios and the segment's international businesses. Tax, Audit & Accounting organic revenues grew 11%, driven by UltraTax, our Latin American business and CoCounsel.

Reuters organic revenues rose 5%, driven by growth in the agency business and our contract with LSEG. Lastly, Global Print organic revenues declined 6% year-on-year. In summary, we're pleased with our Q4 results. Full year organic revenues grew 7%. Organic recurring and transactional revenue grew 9% and 4%, respectively, while Print revenues declined 5%. Adjusted EBITDA increased 6% to $2.9 billion, yielding a margin of 39.2%. Adjusted earnings per share for the year was $3.92 compared with $3.77 per share in the prior year.

Let me finish on the results for the full year by noting that we have met or exceeded all of our 2025 guidance metrics with the loan exception of interest expense. This was higher than our forecast due to the pace of our share repurchase program and declines in market interest rates, which led to lower interest income.

As you know, we have talked a lot about AI from a product and an innovation perspective in recent years. I'd like to close my comments today by briefly discussing another important AI-driven initiative here at Thomson Reuters, namely leveraging the technology internally to reimagine how we work.

Since 2023, we have invested heavily in AI tools and training. We created Open Arena, internal AI platform providing all employees with access to leading AI models and capabilities, and we are leveraging multiple third-party AI applications. We have held several TR-wide AI-focused learning days, offered AI certification programs and taken steps to encourage and foster a culture of experimentation with a test-and-learn approach. This enabled our employees to not only build skills, but also test the impact created by the AI tools.

Last year, having matured from experimentation to execution, we implemented a disciplined top-down approach to driving transformation at scale through AI-driven automation. Our AI initiatives are now purpose-based with specific business targets, and we have created 3-year AI road maps for all segments and functions. We are leveraging key talent, process and learnings from our successful execution of the change program to reimagine how we use technology, simplify complexity, optimize our footprint and commercial systems and evolve our work with the tech plus talent mindset.

We are pleased with our progress to date as we are seeing broad engagement and a growing number of successes. For example, more than 85% of our employees are active users of Open Arena and our other AI tools. We have more than 300 AI use cases in development across all areas of the company.

Let me share four specific examples where AI is already driving measurable business impact at Thomson Reuters. Let's start with software development. Over 80% of our engineers actively use AI-powered tools, making AI a core part of our development life cycle. This goes beyond cogeneration. AI is accelerating security remediation, modernizing legacy systems, automating quality assurance and enabling predictive incident management. These capabilities have transformed how we build and deliver products, reducing lead times and improving quality at scale.

Moving to customer support. By pairing third-party AI tools with internal development, we have automated knowledge search and content creation and added agent assist chatbots. These tools are accelerating routine work and improving outcomes for both our customer service agents and our customers. This includes having reduced call average handle time by 15% and boosted first call resolution by 10%. In content operations, we're leveraging proprietary AI tools to extend our differentiation and content advantage. Our content creation process requires deep legal expertise and proprietary technology.

Over the past 18 months, we've developed specialized AI tools specifically engineered for legal content processing. These tools now automate complex tasks like document ingestion, classification and citation conversion for judicial cases and legislative updates. These advancements improve our speed and quality of proprietary content we deliver. For example, we have accelerated US content delivery to Westlaw by 25%, giving customers faster access to critical legal updates. And importantly, the increased automation also frees up resources that we are refocusing on driving further product enhancement and accelerated innovation.

And finally, our General Counsel's office continues to excel in leveraging Thomson Reuters AI-enabled legal tools. They were early adopters of CoCounsel and our other AI-enabled offerings, which they leverage for contract review and drafting legal research and more.

I'll provide one concrete example with CoCounsel tabular analysis, a recently launched feature that automates the review of up to 10,000 documents at a time. Following the close of an acquisition last year, the team used this capability to review thousands of customer contents, completing this complex task in hours rather than weeks and saving significant time and resources. It also allowed for a much quicker assessment of the ongoing obligations and risk profile of the acquired business. Mike will share some incremental commentary, but let me conclude by stating we have growing confidence that this reimagination of how we work will bring significant benefits to TR over the next few years, increasing our productivity, accelerating our decision-making and improving our customer experiences.

I'll now turn it over to Mike to review our financial performance.

Thanks, Steve. Thanks again for joining us today. As a reminder, I will talk to revenue growth before currency and on an organic basis. Let me start by discussing the fourth quarter revenue performance for our Big 3 segment.

Organic revenue grew 9% in the fourth quarter, continuing the strong trend from recent periods. Legal Professionals organic revenue grew 9% again this quarter despite slower growth from government as we discussed last quarter. Key drivers from a product perspective remain Westlaw, CoCounsel and Practical Law. Government grew 5% in the quarter, though we expect this to slow in Q1 based on the government cancellations we discussed last quarter.

Our Corporate segment grew 9% organically, driven by 9% recurring and 7% transactional growth. Indirect Tax, Practical Law, Pagero and our international businesses were key contributors. Tax, Audit & Accounting organic revenue increased 11%. Recurring and transactional revenues grew 12% and 3%, respectively. Our Latin America business, UltraTax, CoCounsel for tax and audit and SafeSend were key drivers.

Moving to Reuters. Organic revenue rose 5% for the quarter driven primarily by growth from the news agreement with the Data & Analytics business of LSEG and $5 million of generative AI-related transactional content licensing revenue and our agency business.

Finally, Global Print revenues decreased 6% on an organic basis. On a consolidated basis, fourth quarter organic revenues increased 7%. At the end of Q4, the percent of our annualized contract value or ACV from products that are GenAI-enabled was 28%, up from 24% last quarter.

Turning to our profitability. Adjusted EBITDA for the Big 3 segments was $709 million, up 9% from the prior year period with the margin rising 130 basis points to 43%. Reuters adjusted EBITDA was $48 million with a margin of 21%. Global Print's adjusted EBITDA was $54 million with a margin of 39.6%. In aggregate, total company adjusted EBITDA was $777 million, an 8% increase versus Q4 2024, reflecting a 110 basis point margin increase to 38.7% despite $19 million of severance expense related to our initiatives to reimagine how we work.

Turning to earnings per share. Adjusted EPS was $1.07 for the quarter versus $1.01 in the prior year period. Currency had a $0.01 negative impact on adjusted EPS in the quarter. Let me now turn to our free cash flow. For the full year 2025, our free cash flow was $1.95 billion, slightly ahead of our approximately $1.9 billion outlook. EBITDA growth was the primary driver of the year-over-year increase in free cash flow. I will now provide an update on our capital structure and several capital allocation items.

From a liquidity and capital structure standpoint, we remain in an enviable position with below target leverage and healthy cash flow. This strong financial position is illustrated by our December 31 capitalization. We had approximately $500 million of cash on hand at year-end. We have an undrawn $2 billion revolving credit facility. We also have $1.7 billion available for issuance under our Commercial Paper program. Our December 31 leverage ratio was 0.6 times, below our 2.5times internal target, as noted in our value creation model. Looking forward, we forecast $11 billion of capital capacity through 2028.

We remain focused on value creation, and we expect to continue with our balanced capital allocation approach that includes annual dividend growth, strategic M&A and capital returns. I will provide several updates. This morning, we announced a 10% increase in our annual dividend for 2026 to $2.62 per share, up $0.24 from $2.38 in 2025. This marks the 33rd consecutive year of annual dividend increases, and the fifth consecutive 10% increase. In 2025, we completed four acquisitions for approximately $850 million. SafeSend and Additive brought key capabilities to our Tax, Audit & Accounting segment with small bolt-ons of TimeBase and Imagn Images bolstering our Legal business in Australia and Reuters, respectively.

In October, we completed the $1 billion share repurchase program or NCIB, we announced in August retiring 6 million shares. For the year, we returned slightly more than 100% of our 2025 free cash flow through dividends and buybacks, ahead of our 75% return commitment. Looking forward, we have ample capacity to continue to execute against all three of these capital allocation strategies in 2026 and beyond. We remain highly focused on strategic M&A, but expect to again deliver to the 75% free cash flow return commitment in 2026.

I will conclude with a few thoughts on our outlook. We are reiterating the 2026 financial framework we discussed last quarter, and providing additional detail with our 2026 guidance. We forecast total and organic revenue growth in a range of 7.5% to 8%, driven by approximately 9.5% growth for the Big 3. Our outlook calls for modest organic revenue growth acceleration for both total TR and the Big 3.

We are confident in delivering this improvement, which benefits from positive underlying momentum, the execution of our innovation road maps and to a lesser extent, easier comparisons in several areas including at Reuters and Corporates. We continue to forecast our adjusted EBITDA margin in 2026, rising by 100 basis points from 39.2% in 2025. We expect our 2026 effective tax rate to be approximately 19% with our cash tax rate roughly 5% below this book tax rate.

Moving to capital intensity. We see 2026 accrued CapEx as a percent of revenue of approximately 8%, in line with the trend in recent years. This level of investment is supportive of our continued focus on investing in product innovation, as we strive to deliver stronger revenue growth. We expect 2026 free cash flow to be approximately $2.1 billion, up from $1.95 billion in 2025 as growing profitability and stable capital intensity more than offset higher cash taxes.

We are reiterating previously provided 2026 organic revenue growth targets for the Big 3 segments as follows: Legal Professionals growth of 8% to 9%; Corporates of 9% to 11%; and Tax, Audit & Accounting Professionals of 11% to 13%. As it relates to Steve's discussion of leveraging AI to reimagine how we work, I will add a few comments. First, as I stated earlier, our Q4 results included approximately $19 million of severance as we take steps to drive automation and productivity throughout the company. We expect an additional $10 million in the first quarter. While it is early in our automation drive, we have growing confidence in the opportunity ahead.

Combining our underlying operating leverage with expected productivity gains, we now expect to deliver 100 basis points of annual EBITDA margin expansion, not only in 2026, but also in 2027 and 2028 even as we continue to invest in innovation and driving revenue growth. Turning to the first quarter. We expect organic revenue growth of approximately 7% and our adjusted EBITDA margin to be approximately 42%. Looking beyond Q1, we expect revenue growth to strengthen, primarily driven by building momentum from a number of our AI-enabled products.

Let me now turn it back to Gary for questions.

Thanks, Mike. Melinda, we're happy to begin the Q&A.

(Operator Instructions)

Jason Haas, Wells Fargo.

To start, I'm curious if you could comment on the Legal recurring growth. It did slow from 3Q to 4Q, so curious what drove that. And then can you just outline what sort of government headwinds you're expecting in 2026?

Sure, Jason. Your first question, the answer is associated with our Government business. The second question, which is more broader, I will expand. Jason, if you look at Legal Professionals, excluding Government, we have clear momentum driven by the pace of AI-driven innovation including our Westlaw and CoCounsel products. With that said, when we incorporate the full impact of Government cancellations we discussed during Q3, we do see the Government growth slowing in Q1 from the 5% we reported in Q4 of 2025.

We are confident Legal Professionals will achieve 8% to 9% for full year 2026. However, there could vary -- could be variations within that range quarter-by-quarter in 2026 due to the Government. So hopefully, Jason, that helps with both questions. I'll just reemphasize, when we look at Legal Professionals, excluding Government, we have terrific momentum there. It will be offset somewhat with the Government cancellations, but we're very confident in delivering that full year '26 of 8% to 9%.

Mike, if I could just add one more comment, Jason, you might have seen the transactional growth at Legal was really strong. That was actually driven in large part by Government. And so there was a bit of a shift within the quarter from recurring to transactional within the Government segment as well.

Okay. Great. Yes, I was wondering about that, too. So that's very helpful. And then as a follow-up question, I was hoping to take a step back and curious if you could walk us through what is the moat around Westlaw and Practical Law?

And why can't those be easily replicated by some competitor that's using an AI tool to try to do so?

Yes. I'll start that, Jason. Thanks for the question. If you go back to our prior earnings call, we went into some depth with regard to specifically Westlaw. So I want -- I'll try not to repeat too many of those comments, but let me make a few comments because I think this is obviously a hot topic broadly and particularly a hot topic this week.

So the first comment I'd make is we have not seen a change in competitive dynamics in our areas of core franchise. So that's legal research and know-how, Westlaw and Practical Law and the various tax calculation engines that we see and, for that matter, our risk, fraud and compliance data set. So as it pertains to legal, we have not seen new entrants into the legal research space of any measure or importance. Where we have seen more competition is in a white space area for Thomson Reuters, which is this AI-driven workflow area, where the legal assistants have emerged since the rise of generative and now Agentic AI. Now the reason that we are confident and growing in confidence is a couple fold.

Firstly, we believe and all the evidence in the marketplace today supports the fact that a general purpose model cannot do what professional-grade AI can do. So our starting point here, which we believe is a very strong starting point, is that we have two highly differentiated assets. The first, we have decades and in some jurisdictions, centuries of unique content sets that have been created by highly skilled, highly trained and qualified lawyers based on case-law and based on best practices and know-how and improved and refined, curated and delivered over decades or centuries, so the first is our content sets, we think, very difficult to replicate. And the second is our deep domain expertise. So thousands of attorney editors on staff who create this content, and we have retrained to use that expertise to train our agents.

And so as I said in my opening remarks, we have growing confidence here. And that growth in confidence comes from the fact that in August of last year, we launched Westlaw Advantage, which is a deep research Agentic product. Based on early sales and customer feedback, it has reset the bar in the legal research space. To this day, it is unmatched and very significantly more sophisticated, more accurate and more useful than the next best tool. And secondly, in putting this product successfully forward, we have cracked the code, we believe, on leveraging our content and our deep domain expertise to train agents.

And so the opportunity for us going forward and this runs through our product road map in '26 and beyond is to take that leveraging of content and expertise to train agents to our other flagship products, starting with CoCounsel, to launch a fully agentic version in the coming months through to our other franchise products. And so I'd finish by saying there is a marked difference between professional grade AI and general purpose model AI. Professionals -- the stakes for professionals are extraordinarily high. They must be correct. They're doing expert fiduciary work and they require data privacy and security.

And ultimately, they are accountable for being correct. And so that's really a basis of our confidence, not only in our core Westlaw and Practical Law franchises, but also in our ability to extend leadership into AI-driven workflows, which, as I said, is a white space opportunity for us.

(Operator Instructions)

Tim Casey, BMO Capital Markets.

Given the share price action certainly of this week, could you talk a little bit about how share buybacks, return of capital transactions things like that are -- I guess, how are they stacking up in your pecking order? I know you detailed your priority of strategic M&A, but given what's gone this week, how are you thinking about buybacks more specifically?

Yes. Tim, I'll start overall, as you referenced in my prepared remarks, I did mention we will continue with our balanced capital allocation, the three vectors of annual dividend growth, which we increased 10% today, strategic M&A and capital returns. Tim, we agree with your comment and assessment that share repurchases are definitely attractive at the current levels, and we look forward to continuing our ongoing discussions with the Board in that regard.

Lastly, I would mention, Tim, that we plan to deliver on our commitment to return 75% of our free cash flow to investors in 2026. In order to achieve that 75%, we would need about $500 million of share repurchases to achieve this commitment, which is in addition to the dividends previously mentioned. So just to reiterate, Tim, we agree share repurchases are attractive at the current levels, and we'll continue to discuss that option with the Board.

Manav Patnaik, Barclays.

This is Brendan on for Manav. Just wanted to see if you could help us at all with the kind of how the size of the CoCounsel, if that's even possible to split out and kind of the -- it's how much in legal versus tax? And then also, just any comment on the progress of adoption of Westlaw Advantage and kind of what kind of uplift -- adoption uplift you expect that's kind of baked into the guidance?

Yes, Brendan, I'll start. It's Steve. Thanks for the question. So we haven't and don't plan to sort of provide specific dollar ACV numbers for those products as it's relatively early days. What I would say though is that we are very encouraged by the reception of CoCounsel legal by both the general counsels that we serve and aspire to serve and their teams and also the law firms.

And that is both the largest law firms in the world, medium firms and small firms. And I think Steve Assie and Liz Zimick and Aaron Rademacher have done a great job as is Karan Tewari in terms of penetrating CoCounsel into legal community. So we are very pleased with the progress, and we think we are keeping pace or outpace the competition in what is a white space area for us.

And I'd echo those comments as it pertains to CoCounsel for Tax & Audit. On the back of the Materia acquisition and particularly the integration of Kevin Merlini and Lucas Adams from Materia into the TR, the fully Agentic products that they bought with them and have been able to develop with the help of our content and expertise have been very well received by tax and accounting professionals firms, small, medium, large, in-house tax departments and also the professional firms themselves.

So I think it's early days in terms of the development of these (inaudible) change management within the customer set. But we're off to a great start, and we're looking forward to both further investing and innovating and having our fantastic sales team to bring those products and services to market.

Brendan, I think you had a second question in regards to the Westlaw Advantage. I would say the trajectory there is based on actually higher than prior releases of Westlaw. Steve and I had a chance to spend time last week with Mike Dahn. Mike Dahn leads the Westlaw Advantage and also spent time with Emily Colbert. Maybe I'll expand your question which is likely to come, Brendan, in regards to our confidence level in 2026.

These GenAI offerings that we've recently launched is a key ingredient. Several items there, Westlaw Advantage, the CoCounsel Legal that Steve talked about, Steve also talked about Coke Council for Tax & Audit. I would add Ready to Advise, which is within Elizabeth Beastrom's Tax, Audit & Accounting Professionals. And then we just launched Ready to Review with Kevin Merlini there. So that's one of the key reasons, Brendan, why we're confident in our 2026 revenue guidance.

Along with that, Elizabeth is scaling the recent acquisitions of SafeSend and Additive. The last component of that revenue growth acceleration in 2026, I mentioned in the prepared remarks, easier comparisons, specifically for Reuters, given the onetime GenAI content licensing revenue, it was lower in 2025. 2024 was $33 million. 2025, it was $13 million. So that lower Reuters onetime GenAI content licensing revenue, contributes about 30 basis points to total TR if we look at the walk of revenue between 2025 and 2026. Sorry to expand there, Brendan, but I thought that might be helpful just to give you a broader view on our revenue acceleration in 2026.

Drew McReynolds, RBC.

Either for Steve or Mike, with respect to strategic M&A, I think you guys, as a company, clearly have been in AI for a long time and have been successful doing tuck-in M&A. Is it incorrect to assume the pace of all the AI innovation, including from the frontier models is accelerating. So does that make strategic M&A trickier to do? And then when you see a pullback like we've gotten this week across the broader software space, have multiples in the private market -- do you expect them to come in? And just your level of confidence of making sure you're doing the right strategic acquisitions, just given, again, the pace of AI evolution underneath the hood?

Yes, Drew, thanks. It's a great question. Look, in my view, M&A of any stripe and any size is tricky, and it always has been and it always will be. And that's one of the reasons that we've kept, as I think you know, the bar very high. I mean we look for acquisitions that are first and foremost, additive to our customer experience.

Secondly, they've got modern tech. The financials need to work for us, not just the exiting shareholders. And last but not least, we look for executives and teams leading those companies that we think will be additive to sort of TR's culture and talent base. And so we've done a series of these acquisitions, as you've seen. But I think have -- we've kept our powder dry 9 times out of 10 relative to the pipelines and that which we look at.

And I would suggest that will continue because M&A is hard and integration is hard, and we take it very, very seriously. And we're proud of the sort of batting average that we've achieved in terms of integrating the acquisitions we've made over the last 5 or 6 years. As to sort of what happens in the private markets with valuations, as we've shared before, we focus on the best targets and the best products and teams. And so we're not in the market sort of looking for sort of cheap deals and subpar products.

The prices of those assets held in private hands have stayed high in my view. I don't think the private equity firms have sort of readjusted their expectations. Here, we sit with $11 billion of capital to deploy between now and the end of 2028. And I think that's an advantageous position because it should the events, particularly this week, but over the last couple of months in terms of downgrading the valuations of software and AI-related businesses. I think we should be able to take advantage of that, should it occur. But I've been wrong before in terms of how fast the private markets would move. And so we won't overstate the point.

What would you add, Mike?

Good summary. Thank you, Steve.

George Tong, Goldman Sachs.

This is Anna Wu on for George Tong. We are in an environment where many enterprises are keeping headcount flat or even down. How do you monetize your AI product innovations further to capture the increased AI-driven consumptions in pricing. As you look to 2026, how should we think about pricing trends compared to the approach you took in '25?

Yes. Anna, thanks for the question. I'll let Mike answer the pricing question. But in terms of the headcount, so let me make something really clear, which we often state, but I can't be clear enough and that is we do not price on a headcount basis. So we don't sell seat licenses.

So to the extent that our tools and this broader sort of Agentic AI transformation drives greater efficiency and ultimately, headcount reduction within either in-house legal tax audit departments and so forth or the firms that serve corporations. We will be a beneficiary of that, not a victim of that. And the reason we'll be a beneficiary of that, and I used the example in my prepared remarks, of how Norie Campbell's team saved weeks in terms of their due diligence activities and contract review on a potential acquisition is that we price -- we're increasingly looking to price to value.

Now it's early days in this AI journey. And so there's still a sort of a level of experimentation in terms of pricing as customers become accustomed to these tools and start to learn the impact and play through. But we should be a beneficiary of that, not a victim of that. And we're always looking to be -- to base our pricing as much as we can on the ultimate end impact.

Anna, I'll address two additional points I think you raised. One is in regards to how we monitor our GenAI investments. Certainly, Steve just touched on pricing, which I'll go into more detail in a second. Pricing, which we monitor very closely with our segment presidents, Raghu Ramanathan, Laura Clayton McDonnell, Elizabeth Beastrom for the Big 3 segments.

I think another indicator, Anna, is our continued EBITDA margin expansion. As we mentioned and you saw today, EBITDA margin expanded 100 basis points in 2025,and we've made commitments to continue the 100 basis points of margin expansion in '26, '27 and '28. Hopefully, that gives you indicators that we're closely monitoring all of our GenAI and really all investments period. I think that's a strong signal of the return that we're getting.

On your second question on pricing, we estimate a slightly higher overall pricing yield in 2026 versus 2025. And when I say that, that is a composite of our full portfolio, not just the GenAI offerings, but all of our offerings. So slightly higher pricing yield in '26 versus '25, Anna.

Super helpful. Just as a follow-up, I'm wondering what are you hearing from your customers on the rollout of AI solutions in T&A? What's your take on the potential experimenting with other newer or smaller GenAI point solutions that's becoming more available in the market, specifically in areas like invoice extraction or accounts receivable automation versus staying with a scalable and more integrated platform like Thomson Reuters is offering?

Yes. So Anna, your question is, are we hearing customers sort of experiment with general purpose AI solutions in areas like e-invoicing and AP automation? If that's -- if I would answer it correctly, the answer is no. As you know, we acquired Pagero, which is, we believe, world's leading e-invoicing solution and as mandates roll out through Europe and increasingly in the Middle East and Southeast Asia and Latin America, we very much like the trajectory of that product and that offering given its inherent product advantages. And our tax solutions, including our indirect tax solutions and our e-invoicing solutions are lightning fast, very accurate and priced efficiently.

So the idea that sort of a general-purpose AI tool is going to come in and perform that core task more quickly and efficiently at a lower price is not factually true and not the concern. I think the opportunity for us is to use AI as we're doing in the Tax & Accounting space with Ready to Review and Ready to Advise is to use AI to build out the entire value chain of activities around the tax calculation engine. And so that's a playbook that we're rolling out first in Tax, Audit & Accounting Professionals, and we will extend into our ONESOURCE proposition, e-invoicing and indirect tax and ultimately direct tax. And so it's early days for that, but we've got a playbook, and we like the early feedback we've gotten from customers around Ready to Review and Ready to Advise.

Anna, just to supplement to your point, specifically on e-invoice and which Steve commented on. If you look at our indirect tax solution, Pagero solutions, which are included within our corporate segment, those offerings fully met our revenue expectations in 2025 and that trajectory will continue into 2026.

Aravinda Galappatthige, Canaccord Genuity.

Just to go back to sort of, Steve, you comments about the moat. I was wondering if it's worthwhile differentiating the position with respect to your customers and larger law firms as opposed to the smaller law firms. Is there a material difference with respect to the threat from these new generalized services?

And then as my follow-up, perhaps, Mike, on clearly noticed the GenAI enabled products as a percentage of ACV jumped a little bit more than in prior quarters, 28%. How should we think of sort of the high watermark for that? Obviously, there's a bunch of products that will never be AI, Global Print and so forth. Maybe just help us imagine what -- how far that number could go?

Thanks, Aravinda. I'll take the first, Mike, the second. So traditionally, when we put out a new version, for example, of Westlaw or an upgraded version of Practical Law, the largest law firms and the largest general counsel offices were the first to kick the tires on those and ultimately take them up. And then that would sort of work its way down through the customer set. And you've heard over time, Mike and Gary talk about percent of ACV that these products occupy and typically, it reached the limit in the sort of 70-odd percent range.

I think what we're seeing is a slightly different dynamic in this sort of AI-driven era insofar as we have just as much demand in our most advanced AI-driven tools from small firms as we do medium and large because they instantly see the impact in terms of accuracy, quality, and ultimately, the sort of impact to their bottom line. And the small folks see that as quickly. I mean I've been in sales meetings where we've had sole litigators in the Midwest, for example, the US, say, see a demo of the latest Westlaw and/or CoCounsel and say, where do I buy? And when I've asked them, why such a quick decision? They've said because, look, I've been thinking about hiring another paralegal or another associate.

And this does the work in many cases. And it's easier and cheaper for me to do this than go down that hiring path. So that dynamic has been slightly different. But the demand is uniformly keen across all three, which I think is slightly different than previous iterations. Mike?

Aravinda, in regards to your question on GenAI-enabled products, just for everyone's benefit, I'll say context. We've initially provided this metric in November of 2024 with the Q3 '24 call. Initially, it was up 15% of our annual contract value. As noted today, it's now 28%. To specifically address your question, if you can visualize a graph, I think we will see continued steady rise in the percent of ACV that's GenAI-enabled throughout 2026, 2027.

I think the second thing I would do on that visual in regards to the trend line, there will be points in time over '26, '27, where we will add AI to more of our existing products. So at those points, you could see -- begin to see some uptick that's higher than the current trend that we have. So two points there. We'll see -- we'll keep rising steadily. And then as we AI-enable more of our products, you'll see some spikes over the time horizon.

But we're certainly pleased with that continued uptick, Aravinda, and there is a direct correlation to my comments earlier as to my confidence in achieving the 2026 targets with the progress and momentum that we have with the GenAI offerings. Is that helpful, Aravinda?

Absolutely.

Kevin McVeigh, UBS.

Great. Congratulations on the execution and a lot of uncertainty -- perceived uncertainty. I wonder if you could maybe just go into -- because I think one of the parts that maybe isn't fully appreciated by the market is the complexity of the data aggregation. And then ultimately, the overlay of your in-house counsel to draw kind of conclusions. And just trying to get a sense of how difficult it is to aggregate the data, not only from obviously digital but also linear perspective?

So maybe talk to that a little bit, and then I think the importance of, I think, it's at 1,700, 1,800 in-house attorneys you have and how critical that curation process is to the ultimately delivery, Steve, to your point, kind of how important it is getting it right and having the most comprehensive view?

Yes. Thanks, Kevin. It's sort of obviously renewed interest in this question. I would say the vast majority of value in our data sets, particularly but not limited to our lead (inaudible) comes from the IP and the added value from our attorney editors. And these are large teams of terrifically talented, highly dedicated folks spread throughout the world.

And just a couple of sort of anecdotal points. I mean the first is we have in some jurisdiction centuries worth of content and decades and decades of content that was never digitized. So it required someone to go to the court, step to the courthouse and record that first instance. Secondly, to the extent that some of the case information is published by a particular circuit court or a particular courthouse. Oftentimes, it includes a sort of a summary of the judgment.

It doesn't include, for example, dissenting opinions and all of the fact base sitting beneath that. And we've captured that over time. But most importantly is the way our editors interpret that information, quantify that information, categorize it and add the expertise that helps determine, for example, is this patent relevant to other situations and which is the relevant precedent and which isn't. That's where a lot of the real IP is added, and it is a very, very significant added value add over that sort of base level information that's more publicly available. And that is certainly the true in Legal and it's true in Tax & Accounting.

The second thing I would add, I've mentioned this a couple of times, but it's worth reiterating is that under Leann Blanchfield's leadership with help from David Wong and Joel Hron and Emre and many others, we have trained our attorney editors, Tax & Accounting editors to take their expertise not only to produce the content that I just described, but also to train our agents.

So if you think about a particular transaction, legal transaction, we've broken it down into the 25 or 30 or 100 different steps. And then the agents are trained by our experts to behave as a world-class best practice practitioner, and the agents can then behave in sequence and in parallel to work their way through that problem in a very, very advanced way. And so that, I think, is another example of how this human expertise that's been developed here at TR over many, many decades is highly relevant to this AI era and a differentiator.

Doug Arthur, Huber Research.

Yes. Let me ask a numbers question. Mike, the $19 million in severance, a, was that expected? B, where does it show up in the segments? And given the AI productivity displacement, and you mentioned something about $11 million in the first quarter, are we likely to continue to see these kind of small hits as 2026 unfolds?

Yes. Doug, let me hit each of those points. The $19 million, if we go back to the November earnings call, I did foreshadow that we would have some onetime incremental investment. The $19 million is slightly higher than we anticipated at the time of the November earnings call. As I mentioned today, we estimate $10 million of severance in Q1.

If I look for the remainder of the year, I think another $10 million is a reasonable estimate for total Q2 through Q4. That $10 million would be kind of spread along. So directionally, an estimate of $20 million for the full year. Let me bounce back to 2025. You asked about the $19 million and how it's spread.

It is spread among segments and functions throughout total TR. The reason for that, as Steve mentioned during the prepared remarks, we have initiatives across every segment and every function to drive productivity for total TR.

Okay. So it's not in any specific segment per se?

It's not in any specific segment. Once again, Doug, that $19 million is spread among the segments.

Stephanie Price, CIBC.

I'll maybe ask on the EBITDA margin expansion. Mike, it was good to see the '26 preliminary guidance reiterated. Just curious how you're thinking about the pacing of internal AI investments and cost improvements? And what's driving the confidence to kind of extend that margin expansion through to '27 and '28?

Yes, Stephanie. Just to confirm, we did say 100 basis points of expansion in each year, '26, '27 and '28. I think Steve has touched on a few of these points, but I'll expand on it in regards to the confidence level. Stephanie, if you go back to '21 to 2023 when we had our Change Program, we have many members of that leadership team, including Andrew Pearce, Pradeep Lankapalli, Mike Goddard, Liz Bank, who were all heavily involved in the Change Program. So we have leaders who have proved in regards to their ability to galvanize the team, influence, encourage the team to deliver. So that is a really, really important point.

The second point, Stephanie, we have been working on these initiatives for a period of time. They're just not beginning right now. So we go into 2026 with momentum across total TR, as I referenced a second ago with Doug's question. Customer segment, every function in TR with a number of use cases. I think Steve referenced 300 use cases during the course of his prepared remarks.

So once again, those 300 use cases we've been working on throughout 2025. So we're not in exploration mode. We're in execution and delivery mode across all of these 300-plus use cases, Stephanie. In regards to the level of investment, that will happen throughout 2026. All of that is reflected within the guidance. There will not be surprises in regards to incremental investments to actually drive and execute against these initiatives. Did I touch on each of your questions, Stephanie?

You did.

Andrew Steinerman, JPMorgan.

We look at your most sophisticated legal clients in terms of using AI broadly, are these clients consuming more or less content and spending more or less time on Westlaw and why? And I'll just get my follow-up as well. And on my follow-up, I just wanted to get a little more color at a segment level about that first quarter organic revenue guide of 7%.

Yes, Andrew, it's Steve. What a great question. So I don't know that I can give you a sort of the timestamped answer, right, like-for-like as to the talented attorneys spend more time in Westlaw and more time in the AI tools. Certainly, the penetration of these products is growing nicely. And as they are adopted, yes, the time goes up. But like-for-like, I think it's a little too early to tell.

What I would say is a couple of things. The legal profession continues to by and large be in good health in terms of demand for its services. And my own view for what it's worth is that AI will create many, many complex legal issues. And therefore, the sort of headline demand for the profession will continue to grow.

And so within that, these tools are going to play larger and larger roles, and there will be more and more attorneys using them more often each day. And so we think we'll go from a situation where Westlaw would be used by litigators in the context of a particular litigation to a situation where it plays larger and larger roles in the execution of that litigation. And also lawyers are -- one of the first things they do during the day is open up CoCounsel and start using it. And one of the last things they do is sort of shut it down and go home.

So as I've said to you before, we see AI as a means to take our highly differentiated assets and enable us to play a larger role in the success of our customers. And I think what we don't know and what nobody knows is what will the profession's headcount be in serving that. But my bet is the demand for their services will go up, the complexity of questions that general counsels need to cope with will go up, and the adoption of these tools will grow significantly, particularly as they execute over the next 18 to 20 months, the change management required internally, just as we are doing, to ensure that we get -- extract full value from the tools.

Andrew, in regards to your second question, we don't break down our Q1 guidance by segment, consistent with Q1 and prior quarters, I just provided the total. So we don't provide a breakdown by segment.

We have time for one more question.

Toni Kaplan, Morgan Stanley.

I want to put Westlaw aside for a second because I do think that it's easier to conceptualize what the moat is there versus the CoCounsel products. And so I guess with regard to CoCounsel, you're seeing a lot of growth there right now, but there is a lot of innovation in the space from competitors, and I know you're sort of considering them general AI competitors. But I guess if someone is thinking about summarization or drafting or are these like sort of other use cases outside of core legal research, I don't know that it's totally clear why you need the technical expertise for those types of functions.

And so I think that, that's leading to like will the AI product suite eventually either get commoditized or harder to monetize. And so like, I guess, I just wanted to hear your thoughts on how that plays out over time and sort of the risk to growth because of that potential for commoditization?

Yes. Toni, thanks for the question. I think we've touched upon a couple of these points, but let me take a shot at that. I think the first thing you've got to start with in legal is the stakes associated with being a practicing attorney, and the sort of rights and obligations that come with that, the fiduciary responsibilities and the expertise required. And all of those things, the data privacy, the security and ultimately, the accountability for the advice that you provide.

All of those things lead to the need to be correct, and not to allow proprietary information to bleed into an AI model or out into the ether or heaven forbid in the hands of the other side of the litigation or a transaction. And so the stakes are extraordinarily high, firstly. Secondly, the single best way to ensure that the AI output is correct is by training it on a highly curated, trusted, accurate data source. And we have that and very few, if any, of the new entrants have access to such a data source. That's the second part.

I think the third is as we move into this agentic environment and the tools are able to perform more and more tasks and higher sophisticated tasks, whether that be summarization or drafting or production of briefs or rebuttals, motions to dismiss and so forth. In order to perform those tasks, the products need to be agentic. And in order for those agents to behave as a world-class expert would behave and meet the rights and obligations of a practicing attorney, those agents need to be trained by deep experts, and we have them and have invested very heavily over decades in those folks. And as I said in my remarks, have, in a sense, retrained those folks in addition to producing the content also to train the agents.

So that's why we are confident. And we think that our sort of leadership position and heritage in around research and know-how is a great launching pad for us to be a leading player in the workflow tools as well. And we think that more and more customers will look to integrate those particular areas, including research and know-how, and they will want a tool that does both to a very high standard.

Makes sense. And then when you think about the Tax & Accounting business, I think it'd just be helpful. You've done this in the past, but maybe just to hear from your words, how you're thinking about the sort of differentiation or moat there? Because I do think that, that is a little bit of a different animal than legal in terms of what you're providing and differentiating.

Yes. Yes. No, I agree, Toni. I think it is -- there are some similarities, but there are also some differences clearly. So the basis of differentiation is the tax calculation engines that we operate and we provide to customers. And those tend to be very sticky because tax season -- tax filing system seasons come up quickly, and they're incredibly stressful moments for our customers. And so the idea of sort of swapping out the tax calculation engine is not all that attractive for a tax and accounting professional or the head of tax within a corporation. So we operate a number of these products that are leading in their categories, including UltraTax and GoSystem Tax and ONESOURCE.

And so those calculation engines are lightning fast. They are fed by constantly and in many cases, instantly updated content, and they are extremely accurate. And of course, they have the last number of years of tax information embedded in them and feeding through them to ensure that accuracy and relevance to a particular customer. But all of those things make them both difficult to replace and also unattractive to replace because they are very, very high performing.

The opportunity for us is to take that position in tax calculation and extend it into the sort of shoulder activities. So whether that's the collection and ingestion of all the source material that goes into a tax return and SurePrep obviously will do that whether that's the e-filing process that SafeSend performs well or whether it's the sort of integration of these activities that Materia has done so well. And so we think there's an opportunity to take that starting position with the tax calculation engines and to solve a critical problem for the industry, which is an acute talent shortage.

As we've talked about before, Toni, the tax and accounting firms, whether they're the Big 4 or whether they're a high street firm, just can't access the talent as they could 25, 30 years ago, when there were lots of kids coming out of programs wanting to be CPAs. There are far fewer now. So the technology has to fill a significant gap, particularly as the number of returns goes up, the sophistication of returns goes up, the number of audits goes up and the sophistication goes up and the stakes around those things go up. And so that's why we're bullish about both our Tax, Audit & Accounting Professionals business that Elizabeth runs and also the various indirect and direct tax componentry within our Corporates business that Laura runs.

Thanks, everybody. Good day.

Bye.

This concludes today's conference. We thank you for your participation. You may disconnect your lines at this time.