Welcome to Building AI Boston. Our guest today is David Wong, the Chief Product Officer at Thomson Reuters, one of the oldest and most established names in legal information. He leads Co-Counsel, their AI product, which recently hit 1 million users across 107 countries. He also just published a widely read piece in Fortune, laying out his framework for where legal AI is heading. Welcome to the show, David. Thanks, Anna. I'm really happy to be here. Yeah. So my joke is we swim in the same pool, but just a different end of it. Right? So, and what that means is that we're both legal tech people and we're both in this sort of this cool and suddenly very popular space of legal technology, which is interesting. But whereas describe has been around for three years, Thomson Reuters has been around for centuries. So we met just sort of through the LinkedIn universe of talking about all the cool things in legal tech. And I think we were discussing a little bit back and forth, the Fortune article, because we like to nerd out on these things. And one of the funny things I said to David is, everybody knows what Thomson Reuters is. And the truth is, is maybe outside of legal, not everybody does. So, David, give us a primer. Like, who are you and what do you guys do? Why are you so important? Yeah. Well, thanks for having me on the podcast. And I think what is true is that every lawyer knows who Thomson Reuters is, but not everybody I think in the US does. I'm a parent of legal. So I knew exactly who you were and why you were. Lawyer, lawyer, lawyer, lawyer, lawyer, Jason, right? Yes. But yeah, just a little bit about Thomson Reuters. We've been around for over 100 years. So not quite centuries, but over 100 years. And most people know about Thomson Reuters from Reuters, which is their news agency, which is a meaningful part of our business, but actually, this is one of our smaller parts of the business. The vast majority of what we do is provide software and data and information to professionals, professionals like lawyers, like accountants, tax professionals and risk professionals. So the biggest part of our business is legal information and legal software. We provide WESLA, which is the preeminent legal research platform in the US and in Canada, the UK and Australia and a few other countries around the world. And we are, in many ways, the standard for legal research. If you look in the US, lawyers as well as courts and judges all use WESLA as a basis for research for authoritative information on the law and on cases and on any places where you need to do research. And we've also been a leader in the application of technology to legal work. For many, many years, we have invested in and built software for lawyers. And AI is the latest phase of the work that we're doing. And so that's, as you mentioned, co-counsel, which is the legal AI, assistant legal AI platform that we are delivering to our customers and have had in the market for now for almost three or four years. And for the listeners out there, so that's a great primer. And before I got into legal tech, I too knew it through Reuters, right, through the news sort of idea. But there's so much infrastructure and things that are happening in this space that are really important to everybody's life. You just maybe don't know it because it's happening sort of behind the scenes. And the legal space is a really good example of that. And also on the other hand, and David, you're right in the thick of this, it's also presenting some really interesting challenges or sort of opportunities both for legal research or legal tech in general, or the law in general, the fact that AI is coming in hot and strong. Not like we've never seen disruption before from technology, right? Anna has some good stories back from her paralegal days from when new technologies came in. But what do you think people out there should understand? Why legal tech is being so sort of impacted or sort of getting so much attention with the idea of legal, excuse me, with AI coming into the space? Why does it matter so much for us? Yeah, well, I mean, I go back to when Chachi Pt first came on the scene and there was this huge buzz around, okay, what is, what are the industries will be disrupted? And a lot of the research at the time pointed to legal. I remember there was a Golden Sacks article which analyzed the economy just like the economy, everything. And it looked at every single domain of work. And it highlighted that legal was likely one of the industries that would be most impacted by generative AI because it's done on written work product, it's very dense, it requires high amounts of intelligence and of reasoning to be able to conduct the work. But at the same time, it's somewhat rule driven. And you can replicate some of the outcomes with technology and AI. So that was now three years ago, now almost four years ago when that came out. The past three years has all been about realizing if this is true. And in many cases, AI has been an incredibly valuable and incredibly applicable technology to legal work. And a lot of the lessons we've started to learn, the lessons we've learned at Thompson Reuters is that one, the generative AI technology is absolutely well suited to solve many, many legal problems. But it has to be applied appropriately. It has to be applied with care and thought. It's not, it's not straightforward. You can't just take a large language model and say, ah, I've got a solution to a legal problem. You have to design, you have to build systems to solve the particular problems that lawyers are looking for. Because if you don't do that with care and thought, you can create systems which look like they can solve a legal problem, but don't actually perform. There are a lot of cases now with things like hallucination, which highlight the problems of AI systems which are not built for purpose. Can you talk about some of the solutions that you've created using AI for lawyers in particular? I've heard you talk about some solutions that would make me actually want to go back and do law school. I'm actually a real lawyer now. Because full disclosure at the time, I thought, who signs up to do a research paper every day of their life for the rest of time? It's like homework that never ends, but you have some really interesting solutions. And it is kind of funny. Yeah. I mean, the advent of AI has made legal work in many ways a lot more fun, a lot less drudgery goes into the work because of the systems of these tools. Examples where people use chat GPT or use a generic tool to try to do legal work. Those are all examples of the wrong tool for the problem. It's not that AI is not suited. It's just that if you choose to use the wrong tool, you might get the wrong answers. And what we've built at Thompson Reuters are what we believe are some of the right tools. So for example, our latest version of Westlaw, Westlaw Advantage has AI deep research for legal legal research. So we have taken the technology, we've taken the techniques that have been developed for deep research on the web and deep research on other domains, we've pointed it at the legal problem. And this has created a solution which is now arguably the preeminent legal research technology in the world and certainly in the US, we've seen it through adoption, we've seen it through the use. And it solves the problem in a few different ways. The first is that it uses authority, uses information which is authoritative and can provide a factual basis for that research. So instead of going to the web and hoping that the right answer is published somewhere, we have a dataset which encapsulates the US law. It has the history, it has all the statues, it has all the cases, it's all been carefully curated and carefully organized so that the AI agents, the deep research agents, have the information, have the information to be able to answer those questions. So second thing is we have trained these agents to search like a legal researcher. So instead of doing research like you're doing a book report, you're doing research like you were taught in law school to find out, to find precedent to be able to find based off the citation graphs, whether or not it's valid or invalid law to look for contradicting or conflicting cases or information to look at different scopes of or interpretations of the law and all of that's built into the system. So it's the same tech, it's the same underlying sort of LLMs which are which are behind this, but we have pointed it with different instructions and we've pointed it at different information which has helped create a system that can do legal research to be able to again do the work that lawyers would be otherwise doing themselves, but now with the assistance of AI and it's amazing. Our customers are thrilled because it now saves them hours and hours of time. Huffly legal research before AI would often take 20, 24 hours to really kind of fully research a really tough novel issue and now we're saving our customers and we're 50 to 80% of that time. So it obviously is still hard work, but now it's been substantially condensed in legal research. And what the beauty of that is I think in general is that it can just have such an impact on the practice of law, right? Because like you were saying earlier, Anna, like if you had these tools, maybe you would have gone to law school and been a practicing lawyer now because takes away some of the pain and agonies and it's true and you can imagine if you do that at scale, if tools like co-counsel, describe others like help take away the drudgery, all of a sudden there's more room for the art of the law, right? And so people can put more time into that either the human interaction working with their clients. It's very similar like what's happening in medicine. If you've been to the doctor recently, you probably had your doctor coming along with a scribe in tow that's taking notes and things like that. And the things about it that make it unpleasant can start to go away. And also not only for making practicing law a better job, but you're going to get better outcomes, right? Because you're going to have more time to actually spend on the important parts of the process, at least in theory. Oh, no, in practice. Yeah, no, I would chime in here that, you know, most, and I'm glad you compared it to the medical field because I think in cases of doctors and lawyers, people forget there's an oath to do no harm. And when you're, I was practicing in family law, and there is no good outcome. It doesn't matter if you win a case for a divorce client, they're not happy. It's not a happy moment. So anything that streamlines that process and gets them through it, every time a person can, you know, mitigate that lag time or anytime you can speed them through the process, it sounds counterintuitive to a business plan where billable hours are the thing. But I remember at a certain point, someone came up with a way to put, for clients to put their own discovery documents up. And I thought that's brilliant. That's 90% of my work, but clients are so much happier with the advent of what the two of you are doing now, assisting the whole process. I just think it will, A, it will make better cases, better lawyers, it will streamline the process, but also solve a lot of pain points that benefit people in general. So yeah, and I think there is a bit of a trope that, hey, with the increased use of AI, that new risk is getting injected into the legal profession. And I think that is, that would be true if you had everyone just using the wrong tools, right? If everyone said, hey, I'm going to try using, you know, a chat GBT or, you know, a clawed off the shelf to be able to do their legal work, then yes, it would be injecting a lot of risk. But if lawyers are using purpose built tools, right, if they're using tools to design for legal work, I actually think you're lowering the risk, because you're allowing lawyers to be able to find the blind spots. So find the errors to correct issues. As you mentioned, streamline the issues, because we all know that when people are tired and stretched, they make mistakes. And that is, that feels like that's hard for the course in the legal industry, which has ever been as tired and stressed. Right. And we're making mistakes anyway, in terms of, you know, people do as well. Like, and there's a lot of conversation around that, like, well, if you actually compared the error rate, let's call it the error rate, you know, in humans versus in, you know, the machines, like, sometimes the machines are going to actually be doing better. And it's for those very reasons, right, we're human, and we make mistakes, right. So whatever gets to a better outcome. And also, like the really exciting part, I think we all probably think about is more work can get done, which means some of the unmet need out there can start to be addressed at system wide, which would be really good for everybody, right. So David, I know that's an important piece of the pie for Thompson writers as well, the thinking about how how this can be a kind of foundational shift for the whole legal sort of needs out there with all the unmet demand. So, absolutely. I know that that we often talk about the use of these tools within sort of the top end of the market, you know, the most sophisticated lawyers, sort of the Amlaw 100. But we're seeing adoption across the entire industry. So every corner is using this technology. So the courts are using it, judges are using it, lawyers of all sizes and shapes are using the AI technology. And we think that that the ultimate, you know, positive effect will be one, it'll improve access to justice because it'll help to lower the bar for for what it takes to actually hire to be able to to to obtain legal counsel. It'll also just help to make the case backlogs within the courts move faster because they'll be more efficient to be able to to review, to be able to address cases with like, which again, will improve access to justice because of of the time aspect, right, the the the time as you can get get access to to the courts. So I think there's a lot of positive sides to again, the thoughtful adoption of AI tech across legal industry. Well, Anna, you saw it right away, like you used to hone right in on AI and legal, like, that's because that was my background, you know, I'm telling you that part of my diss and, you know, enchantment with the system, I wasn't I worked for a super lawyer, which meant we won cases most of the time. What really disheartened me was the fact that bad actors made outcomes, you know, miserable for people that couldn't afford a super lawyer. And I just thought, this is a terrible system. So when I met Kara in a big room full of people, and I knew what she had, and I was going to interview her on iHeart, I kind of became a cheerleader for you and grafted you into the show like immediately and then out of the gate. I never, I never told you this. We've used Westlaw forever. I mean, we know Thompson Reuters. And that was the that was the word around the office where I used to work was, well, good luck to her. But there will never be a day when she replaces Thompson Reuters. So I want to, so I want to skip ahead to something that's happened in the last week. I would agree with that, by the way. There you go. No, we've replaced ourselves. That's the thing. And that's actually a really important point, which is that if we didn't do that, then we would get replaced. I mean, that's the reason why we're investing in the AI tech, because, frankly, you know, the deep research, the AI version of research on Westlaw is just superior to the old versions. Yeah, if you use if you use Westlaw of five years ago, it's it's it's out of date. It absolutely is. And so Westlaw today is is is so dramatically better. So much so dramatically different. I want to comment on the like, competitive versus collaborative or living in the same ecosystem. So it's really funny, because your your point, Anna, about your about your firm or firm being like, yeah, good luck to her, you know, go ahead have fun with that one. Yeah. And like I say, I agree, because but to enter these spaces, you know, you don't have to necessarily say, and I think this is some mistake that some other startups actually make where they they enter the space, and they say, we're going to dethrone so and so, or we're going to be the new XYZ. And it's, I don't think you have to play like that. I mean, I think AI gives such a different opening to the market. And this is actually kind of what we were talking about in the back and forth on the Fortune article is, there are things you can do in the space and it's what describers doing, describes is doing is where it's you're solving other parts of the ecosystem, you're meeting on that unmet need. You know, like David was saying, like, some of these tools really need to concentrate on certain types of customers. But that leaves a lot of other unmet needs, right. And so there's that's where there's also really interesting things going on is, okay, let the big guys, you know, go up here and worry about the top 20% of the market or whatever it is. And you know, leaves a lot of room for access to justice players for like new entrants and all that. So it doesn't have to be, I think it's a false trade off to say, you either have to dethrone one of the the giant established well loved companies, or you have to just go home, like, yeah, I mean, I think that the the world and the customer need is so large, that there, there's enough space for lots of different companies to create solutions. Even if the competition is high in every single one of those spaces, even if you look within legal and just not even talking about talking about AnthropiGET, there are a lot of great solutions that have been built for particular problems for different types of customers, different types of firms, you look at, let's say, the domain of intellectual property, there's a lot of specific, specific tools that have been built for patent analysis and for patent drafting. And it's, they're better than if you use the generic tool. And similarly, there are platforms who've been developed for like personal injury, because of the repetitive nature of personal injury cases that have been designed just to make it simpler and easier for firms that are specialized in personal injury. And you know, talk to us, we partner with one of them called Supio, because it didn't make sense for us to, to, to go customize something within co-counsel when there's a great solution which is being built for that, that more, that narrower, narrower use case. So I think there's a lot of space. I think that's number one. I think the second thing, why is it that, like, that, that niche companies or specialist companies have been able to create solutions which differentiate, that are, that are able to, to do the job better than others. And this comes back to my fortune article, which is that AI systems are really good at coming up with plans now, being able to describe how to solve the problem. But AI systems are not always great at actually executing that plan. And by that, I, I look at a lot of the legal use cases that were solved for customers. And just to actually execute the work, you need a few different pieces. You often need to have, as I mentioned before, for WESLA, you need authority. You need to have the facts to be able to properly execute it. So even if you come up with a plan, if you have a reasoning system that can say, oh, this is how you would solve the problem. You do step A and B, C, D, E, F, et cetera. To actually execute those steps, you often need to have authority. So you need to have access to the, to the law. You often need to have access to specific systems and tools. So if documents reside, let's say inside particular third party systems, or use particular bespoke tools for analyzing or for processing that information, you need to have those tools. You often need to have those systems with instruction given by experts to know how, what, what does, let's say evaluating, you know, a, a set of due diligence documents for risks. What does that actually entail, for example? And then you obviously need to have access to, to those AI systems to be able to, to accomplish that work. And what we found is that if you have those different meridians, if you have the information, you have the tools, you have the expertise, then you're more likely to be able to create a system that can actually execute the plan, that can actually complete the work. And that's what we've seen with our systems like co-counsel and Westlaw. And it's one of the reasons why we're, we're working with and why we've partnered with Anthropic to extend CLAWD with our, with their MCPs, because let's go back to like the very beginning when we said, hey, is a risk when lawyers use a generic tool. If you used CLAWD out of the box, disconnected from the, from Thompson Reuters or other legal solutions, then you would absolutely be inserting risk because we've seen examples where if you ask CLAWD to do legal work for you, it'll go search the web, it'll try its best, but it might not actually come back with the right authority, the right guidance, the right, the right procedure. The opportunity that we're helping to fill for our joint customers is to allow CLAWD to delegate to co-counsel, to delegate to Westlaw, to be able to do research, to be able to accomplish work when, when it makes sense, when the task is best suited for, for, for co-counsel. And so we helped to fill in the gap. We helped to fill in the gap where CLAWD can't actually execute the plan, can't actually execute the legal work by making our systems available through an MCP connector. Right. And that's the whole, that's, so that's the exciting news that just happened literally yesterday. Today, we're recording this on May, we're recording this on May 13th, and I don't know how much sleep Dave has gotten in the past 48 hours, probably not much. I know I haven't gotten any. Because again, it shows the, the breadth of, of solutions that are out there. And it's exactly the thing that if you are not tethering, so this is why it's risky. Yes, so many people are using these tools, right, for legal work. But what's risky about it, it's tethered to facts that are sometimes real, and facts that are sometimes elucinated and all this stuff. So this is, I, I try to think of different analogies, but it's like inserting some guardrails or, you know, bumpers on the lane when you're going bowling, right, to just keep you from going astray. And so that's super over simplifying it. But it allows you, I think, David, see what you, what you think, it allows you to get the best out of Claude, which were big Claude fans on this show, like we really like the interface and the, you know, the whole sort of ethos of the company and everything else. And then, but get the best out of the other pieces that are involved, you know, or the other tools. And for Thompson writers, that's pretty obvious, you know, it's bringing this fiduciary level is bringing this extreme, you know, deep knowledge. So you can kind of have the best of both worlds, right. And it's, it's a, it's going to be quite a interesting few days to see how the market reacts, right, David, you probably are hearing a lot of interesting things already. Yeah, absolutely. And I think at the end of the day, when we think about this, it's a story of, of coexistence, because we, we have known that particularly for corporate customers, every corporation is chosen to use a generic, a generic AI tool, they're opting for Microsoft or Google or, or Anthropic or, or open AI solutions, because they have a broad set of, of tasks, which they want to add to help with. It could be anything across their business, right, they have more than just the general counsel's office, right. But what they, what they also want to avoid is a situation where, again, the wrong tool is used for a particular, particular task. And so, right now, within a general counsel's office, they have to know how to context switch between using Claude for some tasks, preparing presentations, doing email, etc. And when to, when to use a legal product like a co-counsel or even one of our competitors. And so this helps to make it easier, where if you ask Claude, it's reasoning and its access to our tools will help to direct the task, to direct the, the work to co-counsel, to Westlaw, to practical law, to all of our different solutions and data to be able to fill in that gap. And so it's happening already. When you have two separate tabs open on your, on your, on your browser, we're helping to make it so that it's a better experience for our joint customers. And I was impressed, I have to say, because so, wait, in leading up to this announcement, obviously, we're all super excited, the people who are involved, maybe the, the small startups more than the big established companies, but, you know, it was a big moment for us and really interested to see who else is going to be there, right, because we didn't know ahead of time. So I was particularly impressed that you, you were there, right, because to me that, Thompson Reuters, because to me that really speaks to you seeing what the mark, where the market's going, what people need and how to really be in the mix where people are working, which I think is great because it's a decision, right, to jump in or not. And, and I think it's really cool that you guys did. I think I think it's really, really innovative. Trying to think about our customers, right, trying to think about our users first. Again, you know, I, I come back to, to the general council who is already inundated with all the tech that's being thrown at them and how can we make it so that their life and their team's life is easier and also help to, again, remove risk. Because, you know, if everyone just started to use Claude without the MCPs, without the connectors, you know, off the shelf, I would argue that, that that would be a much harder change management, a much harder kind of adoption curve than to have co-counsel, Westlaw and practical, etc. integrated with, with Claude. So let me ask you two experts. Is this what you refer to as a data layer like this, that your two interfaces are now adding this data layer that's mitigating that risk because you're just adding in a layer of authority? Is that what that means to the street level person? No, actually, I actually don't think, like, because it's so much more than data, right? It's not just that we're providing the facts, we're actually providing intelligence and we're providing tools, we're providing the capabilities of, of, of co-counsel. So I think of it as providing sub workers or sub sub agents or, you know, delegated AI is more of, more of a way I think about it. Because, you know, when you ask Claude to use these MCPs, it's not just saying, ah, like, let me find some facts. Most of the time, it's going to be, ah, I can't do this task. I need to actually work through these integrations to be able to accomplish the task. So it's much more about please research this, this question end to end or help me with this legal task end to end, end to, to delegate to co-counsel. And it's actually interesting, Anna, because I would have answered that differently because, and I think that actually shows how our tools are different, like, they're much, you know, TRs have much larger sort of end to end workstream solution where described, we're actually approaching it a little differently. So that's the data layer stuff is, is a little more accurate for us. And, and, but what that does is so we're leveraging, we have our platform where we leverage our own processes and all that, like David's saying, but we are leaning into the, the things that Claude does. So our tool exists, you can use it fully in Claude using some of their intelligence layers. So my favorite, okay, this is bad, my joke I'm using lately is we've got the brains, they've got the looks. Let's make you know the rest. So, you know, or change the world, one or the other. But yeah, so it's interesting because there are different approaches, but only companies like a TRs level can provide that end to end sort of complete solution where some of the smaller players like us, we are coming in with very sort of surgical sort of solution. So for us, yes, data layers is the right way to say it with, with intelligence sprinkled on top. But then, you know, it's interesting because like if you're saying that, that we've got the brains and they've got the looks, sort of that's the joke. Now I'm in Claude's got the looks. And that's what I mean, which is that describes you at the brains. That's why I think it's so much more than a data layer because the brain is more than just knowledge, right? The brain is also the intelligence you have. And so it's not like you're delegating the thinking to Claude, right? Like you have described doing, doing the thinking as well. I think that's the distinction, which is it's like, it's like having a team of specialists, right? When you, when you use Claude without the MCPs, you are just asking one smart, you know, agent to try to do everything. But instead, what if you had a team of specialists that all can do different types of work, you know, describe as a specialist, we have a specialist, we have many specialists, right, that we're bringing to that team. And now when you ask Claude to do work, it's farming out different tasks, different work to those different specialists. And that helps to complete the work much more completely. So I think we're all providing frankly, intelligence more so much more than just data in these integrations. I think it's just interesting that this is kind of a, again, this is not a plug for Anthropic, but I do, I do feel like can be, we love them. The collaborative spirit is, is, is why I started a podcast in Boston. I just hadn't seen a group of highly intellectual and, and ethics concerned beings that really want to do it right. But you know, here we are connecting with you up in Toronto. David, is that correct? You're in Toronto? Home is Toronto, but I'm on the road today. Oh, okay. Yeah, I just think that what AI is providing more than anything is this human connection and a renaissance rather than a strange inflection point of I don't know what to do with this yet. I just find that this collaborative spirit is really, I hope that we mirror that in other areas, but certainly something as complex as law. I'm, I'm thrilled to see that I hope our audience will take courage in the fact that there are people that are trying to resolve things in the spirit of collaboration. And this is just one great example. Yeah, I think that, that what's also emerging is the ecosystems, software and technology ecosystems, which are evolving, just like when the internet came about and just like when the PC came about, right? You could have assumed that Microsoft, when the PC came out was going to write every piece of software for business, right? And that absolutely was not the case, right? So as capable as, as Excel and as PowerPoint and Word and Outlook and Windows were, there was an ecosystem, which was developed for tools and software and products for every business and every need. And I think that's what we're starting to see now, which is that just as capable and as powerful as Claude is or chat, GPT or Gemini, right? Or Copilot, there's an ecosystem of software and of solutions, which are coming about. And I think that's what we're starting to see this year. Yeah, and it's great to see again, I know, I keep saying this, but I think it's really great to see top operators taking a lead in that, in that positioning and embracing the ecosystem idea. I think that's, that's huge. And I think that's the way of the future. So good for you. Any parting words, either one of you, any more announcements? I mean, I think this is epic, people that are just tuning in and listening. I don't know that you'll catch the drift until you need a lawyer, until you're a lawyer and you see this in practice. But what do you hope for in the future? Anyone want a future cast or take a turn at, you know, where do you hope we go next? Yeah, so I guess my parting thoughts would be, I think we have a really historic opportunity to use these tools to address some of the access to justice problems and sort of fairness issues we've had in our legal system for a really long time, and it had been really hard to change. As David was talking about, judges and courts are using these tools. We know that it's really important for legal aid and for other areas. So I'm just very excited to see how this can finally start closing that gap, because that'll be great for everybody. My parting thought is that, you know, companies like Thompson Reuters, we are going to push the frontier of what we can get AI to do within our domains. And I see the, the opportunities with a cloud with, you know, other players out there as being ways to be able to deliver that intelligence to our customers. That it's, it's a way to provide access to be able to make it easy to use, to be able to convey our intelligence and our information into those systems. And it's not just illegal. It's in tax and it's in corporate compliance and all those different spaces. And so we are going to be shipping some really cool stuff over the next few, few months and years, building agents, which can do more and more professional work. We have some really cool announcements coming up soon around our tax agents, which can help our customers to complete tax returns and to help with tax advisory. Again, stuff that, that brings together the knowledge, the tools, the expertise of Thompson Reuters into our AI systems. And so we're really excited to continue to push the frontier there. And we can get that in the hands of every corporation, every user and every professional out there. Then that's a, that's a huge win for us. A huge win for the world. And I appreciate the high level conversation. I know that people out there, if you, if you want to like and subscribe and hear more, we hope you'll come back, David. And we, we want to do something predictive here and say in a year's time, I would love to see you both again, talking about where this is gone. Sort of the ripple effect. Yeah. But congratulations to both of you. And thanks again for spending time with us, David. It's always a pleasure to hear from you and Cara, well done. And thank you. And yeah, please like and subscribe, audience members, and we'll talk to you again soon. Thank you for joining us on Building AI Boston. Stay tuned for more enlightening episodes that put you at the forefront of the conversations shaping our future.
