Casetext, Inc. (TRI) Earnings Call Transcript & Summary

Good day, everyone, and welcome to the Thomson Reuters Acquisition of Casetext Conference Call. Today's conference is being recorded. At this time, I would like to turn the conference over to Gary Bisbee, Head of Investor Relations. Please go ahead, sir.

Thank you, Paula, and good morning, everyone, and thank you for joining us today to discuss our agreement to acquire Casetext. I'm joined today by our CEO, Steve Hasker; CFO, Mike Eastwood; and Chief Product Officer, David Wong, who will discuss this announcement and take your questions following their remarks. Please note that our intention today is to discuss only the Casetext acquisition and related generative AI strategy and not comment on our second quarter financial performance. We plan to report our Q2 results on August 2. [Operator Instructions] Today's presentation contains forward-looking statements and non-IFRS financial measures. 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. Let me now turn it over to Steve Hasker.

Thank you, Gary, and good morning, everyone. Thanks for joining us today. I'm excited to announce the definitive agreement to acquire Casetext, a leading provider of generative AI-powered tech solutions through its legal AI assistant offering, CoCounsel. We see this acquisition as another step in our 'build, partner and buy' strategy to bring generative AI capabilities to our markets, building upon our organic product road map and the recent Microsoft 365 Copilot announcement. We believe Casetext will accelerate our offerings and our progress and expand our market potential for these new offerings. Before providing a more detailed description of the company and why we're excited about the combination, I'll briefly review the transaction details. Thomson Reuters has entered into a definitive agreement to acquire Casetext for $650 million in cash, which we expect to fund with proceeds from recent sales of our London Stock Exchange stake. Subject to receiving regulatory approvals, we expect the transaction to close in the second half of 2023. While Casetext brings limited initial revenue, we expect it to grow rapidly and be accretive to revenue growth over the next few years. The acquisition initially will be modestly dilutive to margins, though we expect profitability to scale with revenues. We will provide more detailed financial commentary following the close of the transaction. Since its founding in 2013, Casetext has provided solutions that empower legal professionals to work more efficiently and provide higher-quality representation to more clients. Over the past year, the company has refocused its efforts on generative AI and specifically its CoCounsel legal AI assistant offering, which is built on OpenAI's GPT-4 model. Following the March 1 launch of this first-of-its-kind offering, the company has garnered significant industry attention and built an impressive pipeline of customer opportunities. We're excited to have the Casetext team led by CEO, Jake Heller, join Thomson Reuters, and believe that the combination will create significant value for our customers and growth potential for the combined company. Let me now hand it over to our Chief Product Officer, David Wong, who will discuss the CoCounsel offering, the strategic rationale for the acquisition and also our recent Microsoft 365 Copilot announcement.

Thanks, Steve. Casetext's CoCounsel offering is aptly named as it operates as a co-counsel for legal professionals, performing a range of both routine and sophisticated legal-specific tasks, which they call skills. They deliver these skills through a single compelling chat-based Q&A user interface. Attorney users can delegate tasks, including research memo drafting, document review, document summarization, deposition preparation and more to CoCounsel. The automation of these time-consuming tasks provide significant productivity benefits for legal professionals, bringing them up to focus on higher-value work. Casetext has also built a set of internal tools to quickly create these skills. Each skill is a playbook of steps modeled after the workflow of a human attorney or paralegal and is designed to instruct GPT-4 to produce output that is highly relevant and solves for each legal task. CoCounsel currently include 8 legal skills, and Casetext has a pipeline of additional skills in development. We believe CoCounsel is the most advanced legal AI assistant on the market today and see the combination of its leading technology with Thomson Reuters' authoritative and trusted content and insights, providing significant value for the legal industry. The combination of Casetext with Thomson Reuters offers a compelling strategic fit that will provide meaningful benefits for our customers and strong long-term value creation for our shareholders. I'll highlight 4 key investment positives. First, Casetext will accelerate our ability to bring generative AI capabilities to our portfolio and markets. The CoCounsel legal AI assistant has had positive initial industry reception and brings an active pipeline. In addition, the company has created a set of internal tools which can be used by subject matter experts to rapidly create new CoCounsel skills. We look forward to working with the Casetext team to invest in continued development and expansion of CoCounsel, both from its existing road map and also for new CoCounsel skills that leverage Thomson Reuters' content and insights. We believe that these efforts will make CoCounsel more compelling. Second, the acquisition adds attractive new automation offerings to our portfolio that are incremental to our core research capabilities, increasing our addressable market opportunity. 6 of the 8 current CoCounsel skills and several others in development help customers in their work beyond legal research and appeal broadly to attorneys involved in litigation and transactional areas of the law. Third, we see a significant opportunity to create value for our professional markets by more fully integrating our research insights and workflow software solutions. We believe Casetext, along with our Microsoft 365 Copilot integration, are likely to accelerate this merging of insights and workflows, which we believe will bring significant productivity benefits to our customers and a larger position for Thomson Reuters in the legal workflow space. And fourth, we believe Casetext can accelerate our creation and launch of AI assistance in professional domains beyond legal, including our core tax and accounting and risk, fraud and compliance markets. We can use our resources and domain expertise, coupled with the generative AI platform built by Casetext to replicate the concepts and bring value to our broader customer base. Across these 4 points, we see a significant opportunity to leverage our distribution scale and customer relationships to drive upsell and cross-sell of CoCounsel and a growing portfolio of generative AI offerings. In summary, we see a compelling strategic case for the acquisition of Casetext and look forward to working with their talented team to drive value for our customers through an accelerated integration of generative AI across our portfolio end markets. Let me also briefly discuss our May 23 announcement of an intelligent drafting solution that will be offered through integration with Microsoft 365 Copilot. We are very excited about the potential of this offering, which will provide access to Thomson Reuters' trusted knowledge, content and AI technology right where work happens in Microsoft Word. Legal professionals will now be able to leverage Thomson Reuters' content and AI solutions when drafting documents in Microsoft Word and through our plug-in, seamlessly incorporate the authoritative content, analytics and expertise of Westlaw Precision, Practical Law Dynamic and Document Intelligence right into their workflow. This combination offers a compelling step forward in legal drafting automation that we believe will bring significant and lasting productivity improvement to attorneys and the legal industry. It's worth noting that we have a modest position in legal drafting today, and we believe the power of generative AI solutions will expand this product category and thus believe this offering has potential to be an incremental driver for our legal business in the future. Customer feedback to date have been extremely positive, and we look forward to working with Microsoft to deliver this compelling drafting solution to our legal customers as they roll out the 365 Copilot offering over the next few quarters. Let me turn it back to Steve for some closing remarks.

Thanks, David. Let me close by taking a step back and reviewing our broader generative AI strategy. As we discussed on our first quarter earnings call, we see significant potential for incorporating generative AI across our portfolio. We have a long history as a leader in leveraging emerging technologies, including machine learning and AI, for the benefit of our markets. We see a clear path to enhance value for our customers across legal, tax and risk markets by integrating generative AI into our offerings. To this end, we have committed to and are investing to build generative AI capabilities into key research franchises, Westlaw and Practical Law, by year-end. It would bring significant advantages that give us confidence we can be a leader in leveraging generative AI for the benefit of our customers. And we believe generative AI provides us an opportunity to extend our leadership in legal research more deeply into the fast-growing subsegment of legal workflow software. We have committed to investing more than $100 million annually on AI technologies to drive our organic product road map. And as this announcement indicates, we will also pursue strategic M&A that bolsters our capabilities and accelerates our road map. We are confident that our 'build, partner, buy' investment approach positions us to lead the way forward with generative AI capabilities for the benefit of our customers and shareholders. Let me now turn it back to Gary for questions.

Thank you. Paula, we're ready to begin the Q&A.

[Operator Instructions] And first, we'll go to Aravinda Galappatthige with Canaccord Genuity.

Congrats on the acquisition. A couple of -- I guess one question for me. With respect to CoCounsel, I was wondering if you can expand a little bit on what improvements you can drive to the product given sort of the resources that you bring to the table. I know that in the past, that certainly other products used a large number of attorneys to improve the quality of annotation and so forth. What can you perhaps do to -- I mean we fully understand the distribution side, but what can you do to sort of enhance the robustness of the product once it's under the TRI umbrella?

This is David Wong. Great question. I think there are 3 different ways that we're looking at enhancing the CoCounsel solution. Number one is looking to incorporate Thomson Reuters' content. And so Casetext has had access to a research database for some time, but we believe that we will bring a superior set of contents and data with better annotation and better referenceability for our customers that will increase the confidence in their solution, particularly when used in the higher end of the market. I think number two is we will help them to scale. They are, again, serving a relatively small number of customers today, and we have a significantly larger customer base. And our teams are focused on how do we create solutions that, again, can scale to a much larger customer base while maintaining cost efficiency. And then the third is interconnection with software products. The work that we've been doing with Microsoft highlights our partnership strategy in being able to integrate through APIs with software products. And we're looking to do the same with CoCounsel to more directly integrate them with our own tools as well as the ecosystem of legal software products.

And moving on, we'll go to Kevin McVeigh with Credit Suisse.

Great. And congratulations on what looks like a really, really nice transaction. Steve or Mike, is there any way to think about -- as the kind of 8 legal skills, it sounds like there's 8 and more to be developed. But as they scale across Thomson's existing client base, maybe the rate of adoption, and -- does it help the mid-market? I'd imagine it helps the upper marker, too. But how should we think about that? And maybe that within the context of -- it seems like you're in a really nice position, particularly given the technology transformation that we're on the back end of now. But just maybe as we think about it scaling across the existing ecosystem would be helpful.

Yes. Kevin, thanks. It's Steve. So I think there's a couple of sort of ways to think about extending, accelerating our efforts in generative AI and Casetext's existing efforts. The first is distribution. So we think we can bring CoCounsel at an accelerated rate to the sort of larger global large law firms. And we also think we can do the same within the mid -- the sort of what we call the midsized law firms. Casetext have traditionally had a solid business providing content and information to the small law firms. And we obviously have a pretty strong presence there with a variety of products as well. So across that sort of full list of law firms, we think there's an opportunity to enhance distribution of their existing offerings in CoCounsel. We also have started to scratch the surface under Laura Clayton's leadership with general counsels and the broader corporates market -- legal side of the corporates market. We see an opportunity there. So I think that's one way to think about the sort of growth vectors. The other way to think about them is expanding the size of the TAM. And we enjoy -- thanks to the skill of our teams, particularly Mike Dane and Emily Colbert and their folks, leadership positions in the research space, Practical Law, Westlaw and so forth. As we look at this, and David talked a bit about this in his remarks, we're pretty excited about extending that leadership into workflow software and providing solutions to enable lawyers to focus on more value-added tasks, to increase the accuracy and efficiency of everything they do. And so that's, we think, a pretty significant opportunity for us to accelerate that expansion into workflow software and to increase the addressable markets that we play in.

That's super helpful. And then just maybe real quick, any thoughts on the competitive environment? Is it bringing kind of a new set of competitors? Or is it still the traditional firms that you compete against?

Yes. I mean, I think, Kevin, it's -- my view is it's too early to tell. We're very proud of the sort of leadership positions we've established with -- and Practical Law dynamic. I think in the generative AI space, there've been a sort of a series of announcements and there will no doubt continue to be. But my take is too early to tell. I think what we're committing to is accelerating our efforts. And as part of this 'build, partner, buy' approach, on the back of our announcement to invest [ $100 million ] dollars a year, our announcement around the Copilot with Microsoft. This is another sort of data point for you as we think about that journey. And we're committed to establish a leadership position in and around generative AI as it's applied to the legal profession, tax and accounting profession and the risk, fraud and compliance profession. That's very much where we're focused. And we'll also explore opportunities where we see them to exploit generative AI in our news agency as well.

Congrats again.

Thanks, Kevin.

And next, we'll go to Scott Fletcher with CIBC.

I wanted to follow up on the scaling of the distribution question. I'm just curious, is the Casetext and the CoCounsel product as it currently exists, ready to be scaled up and introduced to the sort of large law firms? Or is there work that needs to be done to get it ready for higher -- to higher end of the market for distribution?

Thanks, Scott. Great question. Number one is there's, again, a high degree of interest from all segments of the market, including the top end of the market, with the existing CoCounsel solution. And we believe that for, again, the short term, that they have the capacity, they have the scale to deliver. In the long term, again, we are hoping to be able to address the much fuller set of our market and our customer base, which will require, again, further efficiency and just scaling of their platform. But again, in the short term, again, we're seeing interest from the top end of the market just as, again, as they've had success in the small end of the market. And we believe that they have the capacity to deliver on that pipeline with their existing product.

And I'd just add, Scott, that the access to our content and databases will provide, I think, some further wind in the sales for CoCounsel, and that's certainly the way that the Casetext leadership team are looking at this.

And moving on, we'll go to Heather Balsky, Bank of America.

I'm curious, when you talk about 'build, partner, buy' and the decision to do this acquisition, why buy versus build? I guess, I'm curious about their technology and why you decided to not replicate it yourself and rather buy it and how that kind of shapes your investment plan going forward in terms of the areas where you will invest.

Yes. Thanks, Heather. This is Steve. So as I said in my remarks, we're committed to a 'build, partner and buy' strategy. So the $100 million we announced a couple of weeks ago is very much centered on the build component. And we are -- David and Shawn Malhotra, our Head of Engineering, their teams are -- have moved into overdrive in terms of our own build activities. We will pilot generative AI, integrate it into Westlaw and Practical Law this summer, and we expect to launch in -- generally available by the end of the year. And that's very much along the lines of the build. In terms of partner, obviously, we started with the Microsoft partnership. And as David said, we're very excited about the potential for that to really play a transformative role in the legal profession and make lawyers' lives more efficient and candidly more enjoyable. And then last but not least, the Casetext acquisition. As we looked at Casetext, it very much is a case of expanding and accelerating our own activities. And so we got to know Jake Heller and his team. We got to -- a sense for what they've built in terms of CoCounsel. And so we saw it as additive to our existing efforts. It doesn't necessarily replace anything that we're already doing or planning to do, but it is an accelerant and it is an addition. Anything to add, David?

No, I think the only thing to add is, and this is, I think, well known, that Casetext is one of a very small number of customers -- or companies rather that had early access to GPT-4 from OpenAI. And so they have know-how and experience which is unique in the marketplace. And we felt that, that experience as well as, again, their broad approach to cover a wide range of legal skills was, again, additive to our own strategy to build our generative AI capabilities into our research products and into our own legal workflow products.

I appreciate that. And my follow-up, I know you guys are going to share financials at a later date. But can you provide any color with regards to the, I guess, cash outflow from the business and how it may impact free cash flow once you close?

Yes. Heather, it's Mike. We would prefer to wait until closing to go into that level of detail on the overall financials. We'll be very transparent at that time as always. With that said, Heather, we see significant value creation over time as we drive growth with CoCounsel and leverage their capabilities. So we'll go into all the relevant financials at closing as we do in the second half of '23, Heather.

Thanks, Heather.

And next, we'll go to Vince Valentini with TD Cowen.

Try to push on that a little bit more. Can you give us just any sort of concept of how you think about valuation for acquisitions like this, where you're clearly not enough revenue to value on today, no EBITDA, no free cash flow? Do you look out sort of 5 years to some sort of EBITDA level and discount back? Or do you just say, hey, this is strategic, we have to pay, we have to pay or else one of our competitors is going to grab it, and it doesn't really fit any model but it has to be done?

Vince, fair question. So we apply a very disciplined approach to all acquisitions, including Casetext. So we have prepared our traditional DCF models. We have applied all of our financial metrics and hurdles therewith. With Casetext, as similar with other acquisition, we're looking at the midterm/long-term growth potential. We are very optimistic that Casetext can accelerate our organic growth in the midterm and long term. And over that time horizon, the margins will improve and expand and generate free cash flow over that time horizon. So we applied a very consistent approach, Vince, to this acquisition as we've done others. There's just a little bit longer time horizon maybe with it. But as we go into the second half of '23 and in February, we'll provide more details with it. But the key thing I would take away, Vince, is we think it will be an accelerant to our organic growth as we go into '24 and beyond.

Okay. And a quick follow-up. Is this -- do you think this is sort of indicative of the environment we're in for acquisitions? Is there a bit of an arms race, and we should expect to see more of these sort of Casetext and SurePrep type valuation deals that are great over longer periods of time but dilutive short term? Or is this getting to the end of that cycle?

It's hard to say, Vince, in regards to applying an umbrella-type statement like that. We look at each acquisition for its own merit. We identify additional opportunities like SurePrep, like Casetext, we're going to give it a really, really hard look because ultimately, does it add value to Thomson Reuters' shareholders and does it create value for our customers? And we can check the box, certainly with SurePrep and Casetext on that. So we'll continue to be very diligent with each acquisition, Vince, as we move forward. But ultimately, does it create value for our shareholders and does it assist our customers? I think that's -- those are our two North Stars.

Yes. Vince, this is Steve. I'd just add that we -- as Mike said, we bring a lot of rigor to every acquisition we look at. The vast majority of them, we don't do. And SurePrep and now Casetext, a very small number meet the bar that we set out. So it's that sort of rigor, not arrogance, as we bring it. But we're very focused on our Big 3 and acquiring in and around our Big 3 in both accelerating our efforts and increasing the addressable markets that we're going after. And we think that both SurePrep and Casetext do that. And it's not an accident, they're both AI-driven. So we're proud of our history in machine learning, and we're excited about our organic efforts, to Heather's earlier question. But if we see something that can expand and accelerate those efforts and it meets our financial criteria, we'll go after it.

Vince, I would just capstone that, we'll remain very committed to our balanced capital strategy that we've discussed consistently each quarter with that. And just a reminder, this is an and situation for us. We have the capacity between '23 and '25. This is after completing the recent $2 billion return of capital. Just to remind everyone, we estimate $10 billion of capital capacity between '23 to '25, which allows us to continue with the and, meaning additional acquisitions, additional returns of capital and maintaining that annual 10% growth in our dividends.

And next from Barclays, we'll go to Manav Patnaik.

Mike, I understand $650 million in the context of $10 billion capacity is not much, but -- I know you're not giving details, but can you just remind us of some of your hurdle rates or the rigor that you talked about, your [ presence ] to acquisition? Like what are some of the minimums, numbers of things that need to be met for you to justify this?

Yes. We have a vast array of metrics that we apply, Manav, in regards to, certainly, first and foremost, accretive to our organic growth. We focus very much on IRR. We have hurdle rates there. We focus very much on payback. What's the margin profile over the time horizon? What's the free cash flow, EPS impact, accretive dilution? So a vast array of metrics that we apply. And those metrics, as you would expect, have different aspects of short, midterm and long-term attributes, and we try and balance those overall. So we apply a very rigorous financial quantitative approach. But then consistent with one of the questions earlier, there's always an overlay of strategic importance that we apply and the ability to expand our addressable market over the time horizon.

Got it. And then just on the product side, I guess, I mean, Casetext has been around for a while. And from your own disclosure, very little revenue, so they haven't really been much of a threat to you guys. So it sounds like the advantage of the early access to GPT, I suppose -- like I mean, how many years earlier did they have that access that -- which is -- which I guess prompts you to buy that time, I suppose?

Sure. The -- I would highlight 2 things. Number one is that I think in the last 12 months, Casetext has pivoted itself to be essentially almost entirely a generative AI company that nearly all of their internal resources have been put towards building the CoCounsel solution. And so in many ways, the history that they have is less applicable to the way that we think about evaluating the business today. When it comes to early access to GPT-4, they were given access in the beginning, the first half of 2022 in the early stages of development that OpenAI had with GPT-4. And again, I've had a chance to be able to work with OpenAI to understand the model, to be able to design use cases and to become familiar with how best to prompt the model for effective application within their domain.

And moving on, we'll go to Andrew Steinerman with JPMorgan.

It's Andrew. I know this was touched on a little bit in Slide #6. I was wondering how CoCounsel could help innovation of Thomson's Tax & Accounting solutions products, UltraTax, Check, Checkpoint, et cetera.

I'm happy to address that, Andrew. I think we're looking at it in two ways. One is many of the skills that CoCounsel -- or rather, Casetext has created in CoCounsel are relatively general. If you take a look at what they've published, they'll have features such as document summarization, research within broad databases. And so that general approach is something which we believe can be applied to documents as well as workflows outside of legal relatively quickly. Number two is, again, internally, their tools as well as our engineering teams, we believe, can be applied to use cases when given the right subject matter experts. And that is what Thomson Reuters brings. We have subject matter experts that understand the tasks and workflows within tax and accounting, within research, fraud, compliance and corporate compliance and tax. And we anticipate that, that combination of our, again, expertise and know-how, with their technology will allow us to expand beyond just the legal domain. Steve, anything to add?

Yes. Andrew, so I mean, we've obviously had a chance to think pretty thoroughly about the application to the legal profession. And Casetext's ability to sort of expand our addressable markets therein. The -- to David's comments, the thinking around tax and accounting has started, and there's a few areas where we're pretty excited about taking their skills, the skills -- the 8 skills that they've built and applying them to ONESOURCE to UltraTax and to Checkpoint amongst others. And I also wouldn't -- I wouldn't count out the risk, fraud and compliance space, where we look at combining CoCounsel with, for example, the CLEAR data set. We think there's some very interesting applications therein now. I want to be careful about sort of not overpromising. We've got a lot of work to do on each of those, but we do think that the promise in applying CoCounsel for each of the 3 major clusters of products -- and of course, the business of news and the professional journalism will also be influenced, if not transformed, by generative AI. So one of the questions that we'll be asking of the Casetext team is how they might help out our newsrooms as well.

And moving on, we'll go to Tim Casey with BMO.

Could you talk a little bit about the -- how the process of the acquisition unfolded? How did you come across Casetext? Was it a competitive situation? Just if you could help us know how you ended up here and if there was significant competition for the company.

Tim, I -- it's Steve. I can't comment. You'd have to ask the Casetext team about sort of their process and sort of the degree at which they opened up the aperture. Certainly, from our lens, this was a conversation directly between them and us. It wasn't as part of a process, if that's where your question is going. And as we have accelerated our efforts around generative AI and really explored not only the applications to Westlaw and Practical Law, but also the Checkpoint tax and accounting, risk into expanding the legal TAM, we've cast our eye very broadly as to what everyone's doing. And we think there's some -- we think there's a lot to learn from many different players, including Casetext. So that's really where the conversation started. It was sort of part of a learning exercise on our part, so we initiated those conversations. And the more we learned, the more we liked about it. But we -- these are companies you'd expect that we've known for quite a while.

And next, we'll go to Toni Kaplan with Morgan Stanley.

This was asked a couple of different ways already, but what was really the main differentiator that led you to Casetext needs to be the asset, like we need to buy it for $650 million instead of try to replicate it ourselves? Was it the OpenAI relationship? Was it, it would take too much time to replicate it on your own? Did you not have the right people? What -- and then maybe as a follow-up, like what skills are the most differentiating versus other skills out there with other AI solutions?

Yes. I'll let David comment on the specific skills, Toni, but this is Steve. I would say it's a combination of the things you referred to. The fact that Casetext has had a longer runway than just about anyone in working with GPT-4 and have developed a deep expertise across their team, that was -- so the OpenAI relationship and the work that's been undertaken for 12 months or so was very influential. I think the 8 skills that they've built that are -- at least 6 of them are not skills that we have, that we've focused on or not yet. And so we felt strongly that, that was also an asset. We very much like the Casetext team that Jake Heller has built. So it's really a combination of those things that led us to an acquisition rather than a partnership or purely going it alone. But David, anything on the specific skills?

Yes. What I would say is when you look across the broader marketplace of generative AI companies, there's many which have used GPT-4 and many that have actually been just criticized as being a wrapper on top of GPT-4 that they are finding ways to be able to, again, just cleverly prompt, fairly use GPT-4. But what's unique about Casetext and how they have built CoCounsel is they're one of a very few set of companies out there that also have access to data. They had access to the, again, the historical Casetext research data set. And so they had the benefit of both the early access to GPT-4 as well as experience in how to combine it with an authoritative data set, albeit, again, a smaller, less sophisticated one than we have. And I think that, that made them unique versus others in the market that are just using GPT-4 or other LLMs kind of by themselves. That meant that their team, their know-how, their experience could be much more easily, again, translated into our own teams, where we're, again, one of the small number of companies that have a really authoritative data set and the talent and know-how to apply generative AI. So again, just to build on what Steve said. Within the skills, it's hard to say which is -- which, again, is more valued than another because the application of the skills are very different to customer and to use case. So for example, search a database, which is within their skill set is particularly well suited to, again, knowledge management professionals, legal librarians and researchers and again, something that we're excited about because that's something they were uniquely able to build. But many of the other skills they have, like document review summarization and again, contract analysis are really well-suited to transactional use cases. So again, it's sort of a long way of saying we're excited about the whole set because they apply to the full diversity of legal work that's happening and that we think that we'll be able to, again, expand and grow our TAM by looking at that. The other thing that I would say is the more general skills that I referred to before, like just document review and summarization, can be extended more easily to tax and accounting, to RFC, to any sort of document-driven workflow in the marketplace, and so we'll look to expand on some of those skills that they've built.

That's really helpful. And just wanted to ask also, I think there were a couple of questions on hurdle rates and how to evaluate acquisitions. Like I guess, from an investor's perspective, that we don't have all the information that you do, I guess, what are the product milestones that we should be looking for? What are the benchmarks that we should use to measure ultimately whether this was a good deal for you or not?

Yes. Toni, I'll ask David just to comment on some of the product initiatives that we're going to have during the second semester of 2023. I think that's probably a starting point, David. We've done a lot of work, momentum with Westlaw. Steve mentioned a second ago the work Mike Dane is doing on Westlaw, Emily Colbert on Practical Law. Those are probably 2 good milestones to share with, Toni, in the team just in regards to the second semester.

Again, absolutely right that our focus for the second semester is on delivering on our organic road map. What we've announced in the marketplace is the incorporation of generative AI capabilities to Westlaw Precision, Practical Law Dynamic, Checkpoint as well as delivering on our Microsoft 365 Copilot integration. Again, with CoCounsel, in time, once we close, our expectation is that we'll expand and enhance their set of skills and sell the CoCounsel solution, again, in parallel and packaging with our solutions to legal and then beyond eventually to other segments of our business.

And Toni, as you would expect, as we go into the Q2, Q3 earnings calls, gen AI, and this will be a recurring update that we'll provide and as we go into our March 24 Investor Day. So kind of a continuous dialogue over the next 6 to 9 months on those milestones, both product and quantitative.

Next, we'll go to Maher Yaghi with Scotiabank.

I just wanted to confirm, is it because the company was part of the Y Combinator network, that's why they got early access to GPT-4? And I'm trying to figure out, understand the IP position that they have for CoCounsel. How much of that IP is protected rather than potentially rebuilt or can be reconstructed by somebody else? And also, you talked about increasing TAM. Can you give us or quantify what that increase could look like in terms of the long-term opportunity set here through this acquisition?

Maher, if I could just...

Let me just hit a couple of those points. So Casetext were one of a couple of companies in the legal space that were provided very early access to GPT-4. I don't have any specific insights as to sort of why that -- why them and not others, other than the OpenAI folks seem to concentrate on players that were much smaller and niche in nature rather than -- other than the sort of larger participators. In terms of the IP, we'll have more on that in time. What we've been able to do is get a sense for the 8 skills they've built and the way in which they've been built. And we think that there are some competitive advantages there that are sustainable, particularly when combined with our content, just in terms of the first 2 parts of the question.

And David, in regards to the third-party versus increased TAM, we touched on that during the prepared remarks, but maybe expand on 1 or 2 tangible examples that we're focused on.

Sure. And again, the way that we look at this is we have, again, a large addressable market within the research domain, which we've spoken to kind of at -- in many previous conversations related to precision and the expansion of our work there. We see this as being, again, adjacent to that research opportunity. It is the full set of productivity, technology and tools, which the legal industry purchase as well as the potential for, again, the automation of routine tasks and work, which is right now either outsourced or sent to lower-value providers out there, which we see as, again, another opportunity to expand the TAM for a software or technology-driven solution, which is otherwise completed with, again, some of those providers.

Maybe I can approach this question another way. Do you see this product eventually being merged into your own product and sold within those products or continually -- or will continue to exist as a separate product? And how is it priced right now in the marketplace? How do you sell it? How are they selling it? What price are they getting from a typical customer?

So I think it's a little early to tell, as to -- to answer your first question. But we'll be open-minded about both. So incorporating and accelerating our existing product efforts using CoCounsel and what the Casetext team have built and preserving their exploration of, for us, very new markets and new forms of spend within the legal profession. So we'll have more on that when we close, more on that when we talk, as Mike alluded to earlier, with regard to the financials.

And moving on, we'll go to George Tong with Goldman Sachs.

Congrats on the announcement. There are several legal start-ups leveraging generative AI to automate workflows, including Harvey AI, Dero AI, Robin AI and others. Can you discuss what drew you to Casetext as an acquisition candidate over others such as Harvey AI?

Great question. Thank you. I referred to this a little bit in my answer to one of the other questions, which is there are many companies which are using generative AI. Again, I think it speaks to just how early and again, how broad the market will eventually be. But what I think was unique about CoCounsel and Casetext was that they had the combination of both research databases and knowledge in combination with the use of generative AI, which is, again, not the case for many of these start-ups which are out there and is consistent with our position where we again have a robust authoritative data set, knowledge as well as expertise in AI. So there was that similarity in terms of our profile, similarity in terms of our values and our approach, but then also because they had been able to then build the know-how in how to apply generative AI with that head start with GPT-4 in that combination of both the use of generative AI in combination with authoritative research content.

Great. That's helpful. And how long will it take to train Casetext on Thomson's data set? And is this training necessary before you broadly deploy Casetext? Or do you feel comfortable deploying Casetext broadly before training on Thomson's dataset?

Yes. Great question. And I think the -- I'd be very careful about the terminology because I think that what we found is that there is a degree of training, but more important is, again, the furnishing of information from research databases to the large language models, which, again, is -- doesn't require you to retrain the model. It requires, again, the design of prompts and again, the design of furnishing of information as part of context to those large language models. And so in that respect, we believe that we should be able to move quite quickly to integrate the Thomson Reuters' data sets. We did some proofs of concepts during our diligence and again, give us confidence that we'll be able to move quickly in connecting our data sets with their technology.

I think that's the end of the -- yes, end of the queue. So thanks, everyone, very much for your time. And the IR team is around for follow-up if anybody needs it. Thank you.

Thanks, everybody. Bye-bye.

Thank you. And that does conclude today's call. We'd like to thank everyone for their participation. You may now disconnect.