[00:00:00] Richard Tromans, how are you this afternoon? Very good. Very good. Thanks, Ted. Thanks for having me on the show again. You're very welcome. It's always a pleasure, and we always have really good dialogue, so I don't think today's gonna be any different. Um- There's plenty to talk about. That is for sure. There's a lot to talk about, for sure. Uh, I think most of our-- my audience knows you, but for those that don't, you wanna just give a quick introduction? Um, yeah. Hi, everybody. Um, so yeah, thank you very much, Ted, for having me on the show. Um, so I am Richard Tromans. I run Artificial Lawyer, the, uh, legal tech, uh, and AI, uh, news website. I also am the chairman of the Legal Innovators conferences. We've got Legal Innovators California coming up in San Francisco, 10th and 11th of June. Um, be great if you can join us there. But thank you, Ted. Let's get into it. Yeah. Well, there's, um, like you said, there's a, a ton to talk about. We had a, uh, [00:01:00] Anthropic webinar that was-- had more attendees than the entire ILTA ro- global roster of 25,000 people- Wow which is absolutely astounding that there were that many folks interested. Um- Actually, one, one thing worth pointing out there is that unlike in ILTA, where there's probably one lawyer to every 25 technologists, that was probably the other way around with- Mm-hmm ... you know, 25 lawyers to every one technologist. And, and what does that tell us? It tell, tells us that lawyers, despite everything that you hear, the truth is lawyers actually do want to embrace AI, uh, en masse. They just want it to be a positive experience for them, which is completely understandable. If anyone in any industry was confronted with some new technology, uh, that potentially could help them, pote-potentially could destroy their business, well, they'd be concerned. They're looking for reassurance, they're looking for [00:02:00] guidance, they're looking for some truth. Um, and if they get a positive message which sounds believable, they'll go with it, and I think they're going with it. I think they've started to realize that AI is actually gonna make them more wealthy, uh, who are owners. Junior lawyers, I think understandably, are still a bit nervous, and I, I think there's just a lot of curiosity as well. I think a lot of pe- maybe half the people there were just like, "I don't know where this is going, but I need to know." You know what? I think it tells us something el- I think all of those things I agree with, and I think it also tells us, uh, that there's been a huge mind shift in the legal profession around the strategic importance of technology. Uh, five years ago, this never would've happened. I don't care what technology you would have demoed or what vendor, 25-- 20,000 lawyers w- aren't showing up. So, um- Yeah ... I, I think it's a, it, it is a, an evolution in the industry and a [00:03:00] You know, the tech transformation that made its way through almost every other industry a decade ago is finally now here in legal. They've-- This is like one of the last- Mm ... bastions of holdouts from this technology transformation that's been... I, I was at Bank of America in the early 2000s and, you know, saw firsthand the wave of technical transformation that made its way through, um, financial services, and, uh, here we are 20 years later, and it's finally- Wow happening, happening in legal. Yeah, and it's not a surprise. It's no surprise that it took so long because to change the means of production, my, my favorite catchphrase, y- you gotta get into language understanding at, at an incredibly high level of sophistication, and there was just simply was no technology that could do that. I mean, if you think about the early days of, say, Kira and eBrevia, it could find a clause in a document, right? At the time, 10 years ago, we thought that [00:04:00] was like "Star Trek," right? It's like, "My God, so you put this thing in, and it, like, scans it, and then, like, you know, puts into the thingy, and, like, it tells you where to change your control clauses? My God." And it's like, "Oh, it's amazing. You know, it's gonna change the world." I mean, we literally thought that was literally the cutting edge of AI in the legal sector, and we wondered why 25,000 people didn't sign up to our webinars. Yeah. And it, you know, it- That was- Well, and think about how s- how, how flat that curve was, right? That happened- Oh, absolutely ... you know, 15 years ago. Um- Yeah ... and then it's just been a massively steep slope. Mm-hmm. Yeah. And, uh, and, uh, and it only seems to be steepening. Yeah. Well, you know, the thing about... I mean, when I started, right, I s- I wrote my first paper towards the end of 2014, 20- uh, beginning of 2015, as I was moving out of one consultancy and starting my own, about how AI was gonna change the world, [00:05:00] right? And it was all speculative at that point, and I did a piece, which I mentioned today in the story, about how eventually we'd get to a point where AI would be used by two different parties, and the AI would negotiate and game out the contract using game theory. But I wrote that 2014, 2015, and then I started Artificial Lawyer in 2016, and a lot of people were very skeptical, right? Like outright unpleasant in some quarters. But hey, that's life, right? You know, uh, you know, as they say, you know, you know you're on the right track when people start chucking stuff at you. Um, you know, and but the truth is that, I'll be honest with you, even then I could see it. I just, I could feel it, all right? I could absolutely... I mean, it was one of those transformational just like Gestalt truth moments. I just thought, I absolutely can see this. I didn't know quite how we were gonna get there, but I just had faith that the technology would get there, 'cause I'd studied different areas, and we [00:06:00] talked about this before, the Industrial Revolution, you know, look at the history of flight. There's this really good meme that gets circulated on Twitter every couple of months. There's a picture of the Wright brothers' first plane that actually flew powered flight, and then there's a picture of a astronaut on the moon, right? And it says there's only 50 years between these two events or something like that, or is it 60 years between these two events. But when you think about that, two guys wearing, like, flat caps, right, and shoes with buttons on them, right, managed to fly, you know, was it Kitty Hawk, wasn't it? Mm-hmm. Was it called Kitty Hawk? Was Kitty Hawk the area? Kitty Hawk, North Carolina, yep. Yeah. It flew for, what, 100 meters or something. From there to the moon landings in, what, '68, '69- Mm-hmm how long, how much time is that? Uh, I think it's ri- right around 60 years. Yeah. I mean, that's-- I mean, within one person's lifetime. Yeah. In-- And are we gonna see that with AI? Is [00:07:00] ChatGPT Kitty Hawk? Yeah. That is both exhilarating and terrifying at, at the same time. Um, well, let's talk a little bit about- Mm ... a article that you wrote recently, and, um, I'm not sure if it was inspired by, uh, Scott Stevenson's post or- Yes. Scott, no, it's definitely... I mean, I'd been aware that there were some issues, but I'd never really picked up on it. Scott was the one who really kind of banged the drum. Yeah. And, you know, so I, and I have a take on this as a legal tech founder, and s- we measure contracted ARR and ARR v- very carefully and have very, I, I think, I, well, I thought, well-accepted definitions of what those two things mean. So why- Yeah ... don't we start with how- Yeah we define it, that I think is the most conservative, and I honestly think it's the most logical. So when we sign a new [00:08:00] client, the minute that they sign the paper, that's contracted ARR. They are contractually committed to spend for one year, right? We sign three-year deals- Mm-hmm ... but we, we count that first year's revenue as part of contracted ARR. Mm-hmm. Yeah. For revenue recognition, rev rec, we, we can't re- recognize that revenue, it can't hit our books- Mm-hmm ... until we install our system, and the gap between there is, it really depends on the firm. Sometimes they'll, they'll sign and it takes them six months. Um, sometimes it's six weeks. Um, but that gap between when they sign and when we install the application, um, that's the delta between ARR and CAR. Mm-hmm. An invoice goes out the minute that we install. They're live. Th- it-- there's a accounts receivable entry, and it's revenized GAAP revenue. Yeah. Um, but we [00:09:00] recognize one year, right? If it's a three-year deal, we don't add up all three years. And it sounds like that there's been some very loose interpretations of what CARR means, and it sounds like they're, some of these legal tech vendors are adding all three years in our case, and even, um, POCs that don't ha- they don't actually have contracted revenue. They're essentially a POC that if they decide to convert, they agree on the terms. It's almost like an option. Mm-hmm. And they're counting all three years as, as contracted ARR. Well, we, we, we don't-- well, I personally don't know for certain wh- at least on the legal tech side. I mean, Scott, Scott shared quite a few various things, and you need to speak to Scott about this. But I mean, uh, certainly outside of legal tech, he has shown me several examples which are stunning. You just think, "I can't believe you're allowed to get away with that." And people say, "Well, you know, it's not a, it's not an officially defined term. And hey, insiders know what it [00:10:00] all means. And hey, if you're gonna buy that company, you've gotta do due di- due diligence, and therefore you've shown the books." I said, "Well, that's not the point, is it? Think of all the people that you're deceiving. How is that, how is that right? How is that..." You know, I mean, to me, it's very, very simple, right? If you know you're r- I mean, you know, there's little white lies, isn't there? Like, "Oh yes, I really love, you know, your new haircut," right? You know. Okay, that's not gonna change the world. But y- you're telling thousands upon thousands of thousands of people to invest, you know, to join your company, telling people to buy your software, all of that kind of stuff. What on the, what is in effect a misunderstanding at best, let's be polite, it's a misunderstanding. It's not great. On the legal tech side, you need to talk to Scott about that. I have not seen specific evidence. However, let's say certain people have hinted that certain companies were perhaps a little bit slapdash at the beginning of some of their announcements. Um, that all seems to be reined back. [00:11:00] Certain companies, as you know, I did that follow-up story, certain companies are way more transparent than others, both internally and externally. And kudos to all the people who set out how they do things. It was very commendable. Um, but I mean, basically, I think it was just Scott sending a shot across the bows. And it was interesting to see Gary Tan today, uh, which is, uh, Monday the 27th of April. This may get published later. He did a piece at the weekend, Gary Tan, the CEO of Y Combinator. I don't know if it was Scott or someone else who inspired him to write that article, but he did an article. He wrote, he basically shared an article Um, on Twitter, basically stating all the different types of revenue that any startup founder could use publicly and then defining them. And I think that's the key thing is the definition. So what I would say is, I don't think it matters what metric you use, as long as, A, you're completely transparent about it, B, you totally define it, right? And then [00:12:00] C, you just stick with it, and then you've got a longitudinal understanding. So like if you just stick with-- If you say, "Look, we-we're very optimistic and a bit extravagant with the way we do revenue, but we'll explain how we do it. So every time we get any even marginal hint that we're gonna make some revenue, we book it." And you go, "Okay, so you-- at least you're being straight with me. So you've booked-- So you're booking revenue that doesn't exist yet, that may actually not come about because the client may cancel next year." And you say, "Yeah, yeah, that's what we're doing." It's like, "Okay, well, at least you're being straight with me. That's cool. I think that's, I think it's not a very good metric that you're using, but at least you're telling me." And then from that point on, they keep using that. That's cool because we can see and everyone can know that if you look at X company's revenue data, then you gotta take it with a pinch of salt. Another company might be the opposite, might be incredibly conservative, literally like don't mention it unless the, the check's been cashed in the bank and it's cleared, [00:13:00] right? And anything else is just completely ignored. And you might say, well, the, you know, Y company, they are really conservative. You can totally go by everything they say. You know, in fact, if anything, they're being almost too conservative, you know? But I think that's what, that's what it boils down to. It's, it's not, it's not that the metrics themselves are bad in themselves, it's that people are hiding behind acronyms, and that I think is the fundamental problem. Yeah. And you know, the, uh, there are four main stakeholder groups that I can think of that this impacts. Mm-hmm. Employees- Mm-hmm ... investors, customers- Mm-hmm ... and competition. And, uh, there may be more, but those are the ones that, that come to mind. Investors are not being deceived. Investors, like to your point, have- Mm-hmm ... done due diligence. They don't write checks. I've been through the process. Uh, they understand what the numbers [00:14:00] mean. Um, employees, uh, I do think that there is a, um, there is a credibility issue with employees. They do not get the opportunity to do due diligence when they make career decisions on where they wanna go and what their options might be worth. Um, customers also do not have that level of transparency into, hey, you know, these guys... You know, I may, I may place my bet On momentum that a vendor has in a marketplace- Mm-hmm ... because they look like they're, they're, they're going to be the leader based on new, how many people- Mm-hmm ... firms are getting behind them. I know. Um, and then lastly, the competition. Uh, and at the end of the day, as a, as a founder, I could care less about my competition. If they don't have transparency and so, so be it. But I think the first three groups are, are really important for us to consider, which are- Mm-hmm ... employees, investors, and customers. Investors take [00:15:00] care of themselves. Employees and customers should have-- It, it is incumbent upon a legal tech vendor to be straightforward with your customers and your employees. And, uh, and that's not happening today, it sounds like. And I think kudos to Scott for raising the issue, and hopefully we see a, a correction in the marketplace. Yeah. Yeah. I mean, I mean, I mean, let's not labor this one. I mean, I think, you know, between Scott and Gary Tanner and a few other people, points have all been made now, and I totally agree with you. Um, but ju-just to round that one off, I mean, I would just say that what, what kind of surprised me was how many people, and people some of which I've got a lot of respect for, just were basically like, "Meh, this is no big deal." And like maybe from their perspective, they understand how it all works. They understand, they're very good at finance, and they're like, "I don't know why you're making a fuss about this. VCs know all about this. They know how it works. It's just a bit of froth." But it goes back to your point. I mean, I do think there are real world [00:16:00] harms. Yeah. I mean, I don't think there's, there's two ways about it. There's just, that doesn't work right, you know? And I think just because, I mean, I suppose you could argue-- I mean, it's kind of like, it's, it's like all, it's like many, many, many things when you see this particularly in very, very niche industries, you know, where there's like a nudge and a wink and a yeah, yeah, we all, we all know how it works. It doesn't harm anyone. It's just, you know, it's just between us. You know, it could be like, I mean, you could say, you could say like insider trading, right? You know, two people could say, "Well, look, it's not harming anybody. You're, you're making some money, I'm making some money. Who cares? It doesn't really matter." But it's the same kind of logic, isn't it? It's like basically saying, "Just I, I don't see any material harm to anybody else." Right. But I don't buy that. Yeah. I'm with you. I think it's, uh, I'm glad to see it, it, it raised and, um, I'm glad it's, it's come up in conversation. Yep. All right. L-let's talk a little bit about [00:17:00] the, um, the frontier models and, uh, there was a, there was a podcast a week a- week and a half ago, I think it was 10 days ago, where Anthropic brought their legal team and, um, I got about halfway through it. It's still open in a tab on my browser, so I'm only halfway through it, but I, I thought You know, there were some really interesting, um, tidbits in there, and by Anthropic's own estimate- Hmm ... they think that 25% to 40%, um, of in-house spend- Oh, yeah. That was my article. Yes. Tell, tell me- Tell me what that meant. So, so yeah, so aft- so I, I, I, I, I decided not to jump on the bandwagon and do an immediate response to that webinar. I let things filter through, and I kind of ruminated on it for a bit. And then I spent hours doing loads and loads of different prompts and analysis, both using ChatGPT [00:18:00] and Claude to think about the repercussions. And I eventually got to this little table where it was effectively asking, um, what percentage of legal tech spending would be diverted in the next three to five years to Claude if Claude had the Word plugin, the agentic tooling, and the skills tooling available. So-- and it was broken into big law, medium, small, in-house and, um, access to justice. Right. Um, and I was very interested to see that, so Claude said that the amount of legal tech spending that would go over to Claude from traditional legal tech companies would be between 3% and 8%. And ChatGPT, uh, was four point whatever it was, the lat- the latest version, 5.4, whatever, um, said [00:19:00] between something like 5% and 10%. So it wasn't wildly off. It was interesting. I, I do find actually, if you do-- if you're doing anything kind of data-based, it's interesting to run at least two models against a question to see if they're on the same page, and in this one, they actually were very close. They differed more on the in-house bit. Now, on the in-house bit, Claude said it could between 25, could be between 25% and 40% of legal tech spending would get diverted over to Claude, and that also makes sense. Um, the smaller firms as well, it was very high. So what is the-- what's going on here? I think the point is, is that if you're a small law firm or a small in-house legal team, you probably haven't invested much in legal tech yet anyway. You're probably just running off a Word suite, right? Microsoft 365, whatever. You might have a couple of little tools, but nothing fancy. The-- If you've got enterprise license, if you've got, say, 20 people in a law firm or 20 people in an in-house legal team, moving over to a [00:20:00] Claude enterprise license to do basic contract review, right? Isn't a, isn't-- One, because you're not doing thousands and thousands and thousands, so the cost will be relatively low Uh, it can meet your needs. You don't have massive volume either or f- or massive frequency. It's, it kind of makes sense. Also, the deals that you're doing are not like $200 billion deals. You know, you're not gonna get like shot if stuff kind of like gets hallucinated or whatever. Plus, you know, because you're probably doing lots and lots of routine work, you're working in contractual spaces where you've been a million times. Again, I think the oversight aspect is gonna be very, very high. You know, the outputs, you're gonna b- give it a quick once-over. You know, the documents are not gonna be 500 pages, et cetera, et cetera. So it's, it's sca-- the more you scale down, the more a general application or general, uh, LLM like Claude, which is [00:21:00] actually pretty good straight out of the box on some legal tasks, right? If you added just a bit of tooling and e- even a, even a lawyer who was not a techie could create, um, a skill, load up a playbook, load up a couple of templates, all these types of things. You would-- You do not even have to be technologically proficient. I mean, if you can write an email and arrange your diary on your desktop, right, you could use some basic tooling around Claude. It's no more complicated than that, right? So in organizations like that, I think it makes total sense. So I, I really agree with those numbers. If you're a big law firm, it's a completely different thing. It's a very, very different thing. So like, as I pointed out in the article, so first of all, big law and legal tech are symbiotic. They grew up, they literally grew up together. If you go back to the '80s when big law really started to expand, that's when legal tech expanded. The two are literally joined at the hip, right? The, the complexity [00:22:00] of legal work that they do, the risks of the work that they do, and the multi-practice nature of large law firms and of large corporate legal teams, it's on a completely different scale, right? I mean, the legal market has an enormously long tail. It goes right up like that, and it has an enormously long tail. That big law bit, the Fortune 500 companies and the AM Law 100, they're operating on a scale that is like, you know, you know, orders of magnitude indifference to the rest of the market, right? They need tools that they can trust. You've got the IBM factor. So where did that output come from? Oh, it came from Harvey, it came from Legora, it came from LexisNexis, it came from Thomson Reuters. Okay, fair enough. That's probably pretty good, right? Oh, it came from something we knocked out. You know, Bob, Bob in real estate created a workflow using, you know, Claude, and he told me it was all good. And like, well, yeah, well, guess what? Now, you know- [00:23:00] X bank is breathing down our neck because we screwed things up. You know? Um, it's a totally different kettle of fish. Then, of course, you've got the application layer. Now people have said, like there was this thing on LinkedIn the other day. Someone had basically looked-- modeled, um, Harvey, and then they'd said, "Right, using Claude, if you were absolutely bonkers and you wanted to completely rebuild Harvey from scratch just using Claude and all the various tools, this is how you would do it step by step." But it was ridiculous. Like I was saying to, you know, to, to, you know, someone else on another call earlier, it's a bit like saying, "If I gave you like 10,000 tons of, of metal rods, could you build the Eiffel Tower?" Well, yeah, you probably could, but it would probably fall down once you'd built it. And secondly, what's the point? The Eiffel Tower already exists, right? So the-- you get to this, you get to this point where the, the, there's so much tooling, so much application layer, so many great workflows which have been battle tested, that it [00:24:00] literally gets to the point where there are really diminishing returns from trying to build this whole thing yourself. And of course, the thing is, so, you know, so a law firm throws a ton of staff, maybe two or three different tech consultancies at this thing, builds this great big thing, which is not gonna be as smooth or as good as these professionally designed pieces of software, right? 'Cause that's what they do. That's their job, right? These guys are doing it almost as a hobby, like building a kit car, right? Within a couple of months, something changes in the world, you know, the who knows what it will be, and then you've gotta, you've gotta rejig the whole thing. Plus, who's gonna own-- You can't just own. It's not like owning like one little app. You know, like you've got like a little chat bot function which faces out to the clients and it answers very particular questions on a compliance issue. This is like 70 or 80 different workflows, probably more than that, all stuck together. Who's gonna manage all [00:25:00] of that? Well, this is, this is why things called software companies exist, 'cause that's what they do. So I, I, I, I think it's not-- it really, the, the real-- it is actually a very solid, rational, um, approach to why Big Law won't move over to Claude en masse. And I, and I think that it's very interesting that Freshfields has said they will do this, and that means two things, I think. One, either they're actually not really gonna do very much, and this is just an absolutely fantastic piece of marketing, and they've generated tons of interest, but it will actually result in very little. Or they are taking on an enormous responsibility for internal building and then maintenance. I mean, their innovation team, I would imagine, is gonna have to be very large if it's not already. And then secondly, the partners using all these tools are really gonna have to be up to speed You know, because, you know, who-- I mean, th-think of it, right? Think of this [00:26:00] like very large global law firm with like a dozen practice groups, 30 offices operating in like God knows how many languages, doing thousands of deals a month, right, all around the world, and, and this mass of homemade tools filtering through it. I mean, who's managing all of that? They can do it, but I, I would say that's a huge investment of time and effort. And the craziest bit of it is, is why don't you just bring in a, a ready-built platform that's been battle tested, and then if there are bits, if there are gaps that are not provided, well, then build them. And last bit of a jigsaw is this bit. I was talking to a company in Sweden, uh, the other day called Vesons, and they were telling me about how someone had been using their platform and liked it, and had decided that it was missing a capability. They had built that capability and stuck it on top of this software company [00:27:00] and hadn't even bothered to tell the owners of this company. They'd effectively built a new wing to the mansion and hadn't even told the owners, right? And then eventually they told them, and they checked it out and they said, "This is brilliant." And, and to me, that really opens up a really interesting thought, which is maybe we've got to stop thinking about this idea of it's, you know, it's Claude versus Harvey, or it's Claude versus Legora, or it's vibe coders versus SaaS companies, right? What if we just put the two together, stop trying to see this as a competition, right? It-- to me, unless it's way expensive, bring in a platform, you build out what you don't have, or you build on top of what's there, right? You know, it's like you, you-- it's like customizing a car that you already have. So you buy, you know, your Ford, whatever you have in America, like in England, the Ford Mondeo, most mundane car in the world, right? And then you soup it up. You know, and you stick a roof rack on it, and you put [00:28:00] new spoilers and, you know, whatever, right? And I think to me, that seems more logical. I think this, I think the current debate of, well, we're gonna go all in with X or all in with Y, I think that probably will eventually fall by the wayside as people will realize that's not realistic. I think it-- to me, it feels a little bit like when, um, Steve Jobs was at Apple, and they were fighting Bill Gates at Microsoft because they both wanted to control the entire universe, and then they realized they basically had to work together. So like M- Steve Jobs eventually was like, "Okay, right. So Word is gonna have to operate inside my computers." And Steve Jobs was like, "Well, I don't really like Steve Jobs very much," but basically Word is gonna run, half the world is gonna run their, their Word software on a, on an Apple. So, you know, we better work with this guy. And I think that's probably more like how things are gonna go. At the moment, we've got this kind of like [00:29:00] polarized argument. So that's my, that's my two cents. Yeah. So the most rational takes on this topic that I've heard le-leverage a r- data points. So the average enterprise spends 2 to 3% of their operating budget on enterprise software. 2 to 3%, right? So that means they spend 97 or 98% on other things. So let, let's say I'm familiar with banking, uh, financial services. We're Bank of America, and are we gonna go... We're, we're a bank. Are we gonna go rebuild Salesforce or ServiceNow or, uh, Windows operating systems? The answer is no, right? It wouldn't be a dial mover. It would create an outsized operational risk for all the reasons that, that you mention, and it, it, uh, to your point, I think the [00:30:00] augmentation path makes a lot more sense. Augmenting, filling holes in solutions- Or building on top. Building on top ... or build- building on top or building adjacent to. Mm-hmm. Mm-hmm. Absolutely. But, you know, I've, uh, I, I've chuckled when, um... 'Cause I've had, uh, I've had a few prospects say, "Well, are you worried about, um, you know, somebody vibe coding InfoDash?" And it-- I have to laugh because it's, um, so much of the problem, problems that we solved didn't involve writing software. It was figuring out how to do it. It was, "Hey, you know what? I need this Entra authentication mechanism and this little piece of Azure AI Search over here, and then I need an iManage connector over here." Like, AI's not gonna do that for you, right? No. That require, at least today. Yeah. Um, the coding, now if someone pieced all those ideas and, and we had to iterate 20 times before we found the right combination and the right approach- Yeah ... [00:31:00] um, is, is, uh, is somebody gonna wanna go do that at a law firm and vibe code InfoDash instead of buying our product? No. Well, and al- and also, I mean, another point, and I'm, uh, I'll credit, um, Greg Lambert for this 'cause, you know, this is-- and he's head of innovation at a large law firm in Texas. And he said, "Look, you know, it's incredibly hard to, to bring software into a law firm, and it's equally incredibly hard to get it out again." So, you know, it's, it's really once it get, once it, you know, it's a real marathon to get that software in and get real adoption. And once it's got real use and some love from some of the partners, they become dependent on it. You try taking it away from them. Right. I, it's-- I, I think, I think peop- as in many things to do with technology, and I think to some degree we've been unconsciously trained to see this, we see, uh, a pathway going into the future, and we just go, and then we just apply it to the legal world. And it just, it kind of once you actually, you know, get to the reality of the legal world, it just, you know, these things [00:32:00] just break down. But it's like, you know, the billable hour, right? And billable hour will go away largely at some point in the future. Is it gonna happen anytime soon? No, because just because something can go away doesn't mean that it will go away. You know? It's, uh, the, uh, you know, I mean, I do-- I honestly, I think that what will probably happen is every single major law firm in the world will probably have, you know, 10 or 20, you know, uh, enterprise licenses for Claude, and they'll build little tools every now and then, and it-it'll just be part of their tech stack. It'll be their DIY tech stack, the same way that people used to have Neos Logic, right? To knock out a workflow, right? Um, smaller firms will live in it and die in it, and they'll just take it as it comes 'cause it's cheap relatively if, if you keep down the token cost, which it will, 'cause they're doing like two contracts a day, you know, as opposed to 2,000. Yeah. [00:33:00] And that, that's, that's fundamentally... And I, I, I think it will-- I think we shouldn't overestimate where things will go. Um, the only, the only thing that I think that could really radically upset things, and I don't think there's any chance of this anytime soon, this is pure, pure, uh, speculation. It's not even speculation, it's just like a sort of fantasy, uh, scenario, would be if OpenAI or Anthropic decided to go all in on the legal vertical, the same way that they might go on the medical vertical. They go all in, let's say in five years, 10 years, you know. Let's say they've, they've IP... They've both, they've both IPO'd by this point. Global use of AI has saturated. They're thinking, "Oh my God, we can't grow this business anymore. How do we, how do we grow the business? Okay, let-- You know, we've been messing around with legal for years. Let's go, let's do it for real this time. Let's buy LexisNexis. Let's buy Thomson Reuters, the legal division." Right? 'Cause by this point, they're worth a trillion dollars each, right? They can, they c- they [00:34:00] can actually quite easily buy at a good premium either of those, right? So they have the data. They bring in all that additional, uh, client goodwill. By, let's say, five to 10 years, the AI models are going up incremental- incrementally between one and a half percent every year, right? So five to 10 years. They're at about 90, 91 at the moment. They should be approaching 100 in terms of accuracy and overall output across most general benchmarks. Um, yeah. Then things, then maybe all go, all bets are off. Because you, you just buy one single license I see anti-patterns to that. So l-look at what's- Yeah. Like I said, that, that, that is a pure- Yeah ... pure speculation. I've got no idea that that would happen. You know, look at, I think a, a good case study is what's happening right now between OpenAI and Anthropic. Mm-hmm. Like Open- OpenAI got distracted, right? Did. With shiny objects, Johnny [00:35:00] Ive, Sora- Bullshit basically ... you know, all, all, they're-- Yeah. Yeah. Uh, and, and as a result, Anthropic- Mm ... is kicking their ass in, in delivering the best absolute models and technology. And I think that lesson holds true and will hold true in the future where, you know... Look, look at Microsoft in legal. Mi-Microsoft has tried to step in and provide legal specific solutions. I don't know if you remember Matter Center, God, how many years ago? At least 10, maybe 12 years ago. Um, complete failure. Mm. It was a, a, a huge flop, and the people who invested in it, Epona picked up the tech and ended up building a document management system on top of SharePoint with it, but Microsoft completely abandoned it. You know, these, these general technology companies, the legal market, I mean, it's a $1 trillion services market. I [00:36:00] mean, that's, that's not nothing, but when you're looking at the whole broad economy, that's a really- Yeah ... small slice of the overall economy. No, it's, it's still a biggie. It, it is. If you look at various other verticals, it is, it is significant. And the other thing is the legal market does have a lot of complex work. Um, I mean, uh, and also the world's becoming increasingly regulated, you know, and there's always more laws and so forth. Um, I don't know. I, and also there's this unaddressed legal market aspect which, you know, crop- crops up every now and then. You know, people estimate you could double that trillion if you could bring down the cost of legal services to a point where the average person could get involved, in the same way that, you know, medical services were basically out of reach except for the wealthy until, you know, the sort of early 20th century when, you know, sort of improved methodology and, you know, better technology and so forth. And also, like in Europe, for example, you know, [00:37:00] um, you know, tax money went towards supporting like the NHS and so forth, massively expanded the provision of legal, of, um, health services, right? Who knows? But the... I don't know, I don't know, I don't know. I mean, like I said, we do not know. We do not know. We absolutely do not know. I mean, as things currently stand today, and at least for the next four or five years, I think the general shape of the market will actually stay relatively the same Um, I think the Harvey, LawGora, Clio, Thomson Reuters, LexisNexis will just expand outwards on their AI capabilities or draw more and more people in. The point solutions will battle hard, particularly ones in the contract space will battle really hard to avoid, um, getting submerged by Claude. Claude and OpenAI will get better. Um, you know, I think it will stay roughly to some degree, I know it's surprising that I'm even saying this, but I, I think to some degree, you know, I think we, we've got a general shape of the market now. Um, there may be [00:38:00] massive surprises. There may be massive, massive, massive surprises. Maybe OpenAI gets its, uh, game together, comes back 10 times stronger. And, but I think going back to your point, I mean, why would a general software provider go all in on one vertical? That's what it- That's what makes sense I mean, if we boil down everything we've been talking about, right, in general, if we don't get to that, you know, it, it all boils down to this: Will a general tech company go all in on a vertical? Right. That's basically it, isn't it? It is, and I think the answer is no. Right. But, um, I, there's one topic here I wanna make sure we touch on before we run out of time- Mm-hmm and that's the, the AI native firms. Oh, yeah. I love them. And it's, it's really cool to watch and, um, I'm having a blast seeing how the different iterations, you know, you've got Norm and General Legal and [00:39:00] Crosby and, um, Garfield, and you've got all these different models and, um, I'm curious what your take is on now that we're seeing with, with Norm specifically- Mm-hmm you know, you've got Blackstone behind them. They got 50 million in funding. Uh, just Blackstone by itself could sustain, keep them busy for the next 10 years. You know, I, I feel like we're starting to, um, level up. Like- Mm-hmm ... you know, the, you've got the former, I think he was the former managing partner at Sidley who-- or Skadden, um, who's at, at Norm. They've, they've, they've got a lot of very f- I mean, the one that I remember, which I wrote about, was Bill Mone, who is a very, very, very experienced equity partner at Ropes & Gray- Mm-hmm ... who's in private equity work. They, they've hired a whole bunch of other people. They've hired a lot of very, very, very senior people. It'll be interesting to see how that pans out. Um, I [00:40:00] mean, like I s- I was talking to a bunch of different people about this, and you wouldn't-- I mean, 'cause I write a lot about new mods, as I call them. Mm-hmm. 'Cause I think eventually everybody will be an AI-first law firm, so I'm kind of jumping, I'm getting ahead of the curve. I'm trying to by just saying they're new mods 'cause they have a new business model in that they are, they've done the billable hour. Um, the idea, I mean, eventually all legal work will be AI first. You know, even Slaughter and May and Wachtell Lipton eventually, the fir- the first draft of everything will be AI first, right? You know, give it enough time. So in some ways, you could argue AI first doesn't make any sense as a, as a, as a, you know, as a description. Anyway, the, the key point, and as I've written about this several times, um, quite a few people call me up and say, "Hey, um, you know, I've been working in legal tech, and I'm about to start, uh, one of these new mods," or there's a partner who's thinking of leaving or a general counsel who's thinking of leaving. I mean, you wouldn't believe how many people have called me up and basically said, "Could you give me [00:41:00] some tips on how to build one of these things? 'Cause I wanna do it." Right? There's more coming. I can tell you for a fact there are more coming, way more coming, right? Um, but I always say to them, "Look, the tech bit is actually the smallest part of this." Right? The, the tech is, the AI bit is table stakes. If you get a decent CTO and you know what you're doing, you will be fine. You will get over any technical humps very quickly, 'cause what you're doing isn't rocket science, right? You're not pushing the frontier of technology. The complicated bit isn't even the legal bit, because if you've got good lawyers, they know what they're doing, and the, the good lawyers and the good legal tech people will fuse together the same that they do in a law firm. So you say, "Well, what's, what's the difficult bit?" Difficult bit is the business model. How do you make money selling contracts or contract review for $500 a pop when they're high quality for very large corporates? So we're not, we're not doing, we're [00:42:00] not doing like LegalZoom, you know, "Hey, you know, can you make me a one-page employment contract? I only wanna pay $30." And it's like, well, you get what you pay for. Right? This is like Fortune 500 company emails one of these new mods and says, "All right, we're gonna give you a try. If you screw up, if the quality is bad, if we don't like anything, if there are any errors in there, if we don't like your service, you're gone. We never use you again. We've got, we've got a queue of people around the block who wanna service us." Right? But you're only gonna make $500 on that piece of work. You've got all your tech costs, you've got your offices, you've got your marketing budget, uh, and you've got a bunch of lawyers. Huge cost space towards a tiny profit output, right? And yet the ones that I've spoken to, like, you know, General Legal, this is all on record. There was, um... I did a, um, a video with them. They were saying they actually think that they're gonna actually increase their [00:43:00] profit margins as time goes by, 'cause they're gonna get their processes in, in better shape. The tech will get better, and as they do more and more of the same contracts, they're gonna be able to codify and improve their workflows and, you know, and so forth. And so then, then it becomes all about volume, and volume for a services business is about marketing. And it's also about hiring good quality people who are affordable. So you don't want people too expensive for that work, but not too cheap. If you hire a bunch of paralegals who don't know what good is, and they are the last line of defense be-before these contracts get sent back to this bank or private equity fund, and you just do AI plus a not very well-trained lawyer, and then you send it off, well, that's-- your business isn't gonna last very long. You're gonna lose a whole bunch of clients, right? And then how do you manage these people? How much time do you hire very, very expensive [00:44:00] experienced managers? You know, I mean, for me, it's an organizational issue. It's about people, it's about expertise, it's about cost bases, it's about marketing, it's about building a brand that people will trust, it's about client relationships. You're gonna need a ton of people to go out there and do GTM stuff, take people to dinner, take them to the opera, whatever, take them golfing, right? You know, this is a-- and it's so, it's so weird because the amount of people, particularly who come from the tech side of life, are just kind of like, "Well, we'll build these workflows and we'll do this, we'll do this, and then we'll put out some adverts, we'll hire a bunch of junior lawyers, and off we race as we go." Right? And you're like, "Yeah, so what about... Okay, so that's like you've taken two small steps down the road. You've got about another 20 or 30 to go before you've actually got something that's gonna last." Look, look at law firms, right? I mean, people really, you know, I mean, okay, I, I'm always like poking my finger into the, the big law bear, right? 'Cause I've, I've been working next to it for 25 years, so I know it pretty well. But I've got a lot of [00:45:00] respect for big law. The big law law firms are amazing organizations. They are brilliant at marketing. I mean, I heard this amazing story about this one guy who moved his kids into a school simply because that would allow him to socially get to know the parents of the other kids who might be clients of his one day. That's how far they-- that's how far he would go to secure- Yeah ... a client win. He literally sh- changed, changed his kids' school. Yeah. It's so, so much of a, um, enterprise external legal spend is bought through relationships. I, I'm in the middle of writing an article on this, and BTI did a 2025 practice outlook survey where they surveyed 300 plus GCs, and s- 79% plan to bypass [00:46:00] formal RFPs and cherry-pick from referrals and thought leadership, right? That's a pure relationship game. Uh, the question is: how is that gonna change over time? So like today, you know, I'm, I'm using, for those listeners- Mm-hmm ... I'm making a pyramid with a, with a wide base. Mm-hmm. As the pyramid shrinks, as the base of the pyramid shrinks And you have more, um, in the legal ranks, right? This isn't the overall staff, this is your legal capabilities. As that shrinks and you have more data scientists and AI engineers and business analysts and product managers- Mm-hmm ... how does that, how does that change the equation where you don't have-- where the information isn't all in one person's head who's going out there and they're the $4,000 an hour, you know, rainmaker? Um, how does that change the equation? I think that it does, and I think that the way, not only that, that corporate buyers will buy, but also how they will [00:47:00] manage. Mm. They're gonna need to see a dashboard. They're gonna wanna know what's your turn time on these, on these document. What are your quality metrics? Um, w-what are your cost metrics? It's gonna be a very-- like for the operate the company work that really does get delivered largely with tech, that's gonna drastically change how not only they buy, but how they manage. But going back to how they buy, like today, there is a massive disconnect. If you look at law firms' websites, I'm in the middle of doing this as part of the article. I had called Claude Cowork go out to all a- 100 of the Am Law 100, uh, public-facing sites and document their primary and secondary differentiators. Um, it is a very thin- Mm ... layer that talks about tech- Yeah ... and innovation. It's, um, it's, it, y- and the reality is that, that is going to have to change. So as a buyer, if I'm sitting, if I'm sitting in the, I'm the chief legal officer at Bank of [00:48:00] America and I need, you know, somebody to manage my employment matters that are largely gonna, or my immigration applications as I'm bringing H-1Bs in, uh, to staff my business, am I gonna buy the same way? Am I gonna reach out to my network or my peer group and is social proof? Or is it gonna be the, am I gonna be looking at numbers- Hmm ... um, to decide who my legal partner is gonna be for that operate the company work? Do you have any thoughts on that? Well, it, it's a mix. It's a mix. It's a mix. I mean, yeah, I totally agree. You know, I was saying, I mean, relationships rule, you know, culture. You know, as they used to say, you know, culture eats strategy for breakfast. I mean, I remember years and years ago, I kind of like flipped that Peter Drucker quote a little bit and I said, you know, "Strategy eats legal tech for breakfast." Sorry, start again. Culture eats, um, legal tech for breakfast because, yeah, I mean, you know, even I who, you know, massive, massive proponent, proponent of AI, yeah, I mean, you know, look at [00:49:00] myself. I mean, who do I, who do I do business with? I do business with people I like, right? I mean, you know, I mean, there, there are, there are certain companies who've asked if they can sponsor certain things and get involved with Artificial Law, and I've said no. Or I've made absolutely zero effort to reach out to them, or I've not replied to their emails, even though they'd like to get to know me, and I'd just be like, "No, thank you. I met you once before, and I do not wanna know you." Right? Not many. It's a tiny, tiny handful, but still, it's human relationships, right? It's the same with law firms. It's like, "Oh, yeah, yeah, you know, I met you at blah," and, "Yeah, I'm not sending any work to your firm." You know? Right. Yeah. It's as simple as that. I mean, it's like, it's not rocket science, right? And then of course, there are certain things like, you know, yeah, like, you know, this person is literally the best, you know, person in the entire world at this kind of work. They are literally a, a rocket scientist, right? You know, you'd be mad not to send it [00:50:00] to them. You know, no matter what you think, you, you, you've gotta send it to these people. Um, on the operational bit, again, you know, you gotta remember, if you're a large corporation, you're not spending your money. You're spending someone else's money. As long as the costs are not wildly disproportionate compared to what your peer group is spending, you're gonna lose your job. Well, it goes back to the IBM factor and so on and so on. It's like every year you see-- like, you know, I say this a million times. Every year or every quarter, you'll see these surveys, usually from the Big Four or from the CLM companies, saying, "In-house lawyers are under pressure to do more with less." And you're like, "Right, so legal spending will go down this year." They go, "No, no, it's gonna go up." Okay, so that's a contradiction. Um, okay, and the billable hour will be under pressure to drop. "Oh, no, no, no, it'll probably go up by about 10%." Okay, so that's another contradiction. So, the use of ALSPs will massively skyrocket. "No, no, it'll probably stay about the same." So that's another contradiction. So it's just like, come [00:51:00] on, guys. It's like in-house legal teams are not under massive pressure to cut costs. Maybe some are. There'll probably be someone who'll listen to this and go, "That's not true. I'm under massive pressure," but they really are the exception that proves the rule, right? Yeah. Particularly in the US, it-- I wouldn't say it was a blank checkbook. It's not that wild. But, you know, it's a h- super, super legalistic society, right? I mean, people out in America probably don't realize that because obviously you live in America, but, you know, if you move to, say, I don't know, Switzerland, you'd be like, "Wow!" You know? It's like people here just like shake hands on a deal and walk off. You know, it's like that's really, that's it. You know. Um, and it's, yeah, it's, uh, I, why would that stop? It's not gonna stop. It's gonna generate more work, you know, so on. So yeah. Um, again, it, it, I think it all boils down to like moving away from the, from the, the AI first new mods, [00:52:00] whatever it is. I think-- Well, just on that point, I think they will carve out a good chunk of work, but I think we'd be kidding ourselves if we thought that corporate America really is desperately trying to cut costs on legal. I don't think it is. Yeah, that's a-- W- we could dive into that. Um, I had, uh, I had the chief legal officer from DHL on, um, and he has a different take. Hmm. But, you know, I do think that it is, it is enterprise dependent, and I do think it's going to increase. Mm-hmm. I think that as AI starts to get pushed out within the organizations themselves, there's gonna be natural conversations that arise around efficiency that maybe didn't in the past. Um, but we'll have to save that topic 'cause we're almost out of time, and I wanna give you a minute or two to tell us [00:53:00] about, um, your conferences coming up, when they are, where they are, and, uh, and then we can wrap up. All right. Okay, cool. Um, hopefully I'll be back before the, our autumn, our fall, uh, events. But for the summer, we have two events. Uh, we have Legal Innovators California, which is in San Francisco. That's legalinnovatorscalifornia.com. I know it's a long one, or just type in Legal Innovators California into Google. Uh, it's two days, law firm day, in-house day. It's on the 10th and 11th. Um, it's, uh, right in the heart of San Francisco. We've got a ton of great speakers already. We've got people like David Wang from Cooley. We've got top people from Google and other companies coming. I mean, it's gonna be a small guys board of the West Coast tech companies, law firms, and the best of legal tech as well. It's gonna be a pretty awesome event. It is free to attend if you go through the express registration. [00:54:00] Express registration, there's a whole bunch of different stories on artificial. If you go to Artificial Lawyer, search around, you'll find one of these express registration links. If you click on it, if you work at a law firm or if you work in-house inside a large corporate or an interesting corporate anyway, uh, just apply, and if you qualify, you get free entry, um, on both or either days that you might wanna come. Uh, we're doing another event exactly the same in Paris, which is for the European audience and also people from the US and UK who wanna sell into, uh, Europe or, uh, connect to a European law firm or business, and it's the same thing, and that's on J- uh, June 24th and 25th. And in case you're worried, uh, the conference will be in English, uh, although it will be in France. And is, uh, a-are you still doing this in partnership with Cosmonauts? Yep, Cosmonauts are organizing. Um, they do a very good job, and, uh, they'll be on the ground, and they'll be organizing everything. Yeah. Okay. [00:55:00] Awesome. Yeah, we're new sponsors of their program, so, um- Okay. Fantastic ... yeah. And yeah, if you're wondering what my part is, I effectively, uh, me and Timo effectively came up with the idea. Uh, I chair the events, uh, um, all of it. And also we'll be doing two events, one in New York, one in London in November, towards the end of the year. But I'll tell you more about that as we get closer. Awesome. Well, Richard, it was a great conversation, as it always is, and I appreciate you spending a few ti- uh, few minutes with me, especially this was your third podcast of the day. Uh, so you're probably tired of talking, but I appreciate you. No, not a bit. It's, and it's for, it's my favorite subject. It's my favorite subject. Well, um- And it's been a pleasure to talk to you, Ted. Absolutely. And, uh, you will see my team at your Paris event. Uh- Oh, really? Yep. I have, I have two folks coming. I have, uh, someone in Spain and the UK, and- Mm ... they'll be, uh, they'll be joining. So, um- Oh, fantastic ... I'll, I'll tell them to say hello. Fantastic. Awesome. [00:56:00] All right. We'll chat soon. Au revoir. All right. Thanks, Richard. Bye. Thanks for listening to Legal Innovation Spotlight. If you found value in this chat, hit the subscribe button to be notified when we release new episodes. We'd also really appreciate it if you could take a moment to rate us and leave us a review wherever you're listening right now. Your feedback helps us provide you with top-notch content.