#93 - Harvey CEO: Individual Productivity Is a Trap

I don't think the billblower is gonna go away instantly. You're gonna not gonna get like crazy games from AI unless that you change how your team works and then how your company operates as a whole. I think we're gonna get to the point like very quickly where it's actually like very difficult for just like humans to do this work without AI. - Hello everybody and welcome back to another episode of Pearls On Gloves Off. I am your host Mary O'Karrel. All right, this week most recently what are we seeing? Well, there was a call out on both Twitter/X and linked in about legal AI, legal tech companies and are they truly reporting their real revenue run rate numbers? And there was a difference between, let's say what is generally known as ARR versus CARR, which I guess is contracted ARR. And a question of whether these companies are over reporting their success and that call out went out in a bunch of companies including Harvey, LaGora and many others came out very quickly to say that there is nothing to see here and that the difference basically between those numbers is negligible. So perhaps this whole thing was a nothing burger. The other big thing happening this week was companies like all birds doing crazy things like shifting from being a maker of wool shoes to a compute company and their stock price went up like 400% in a day. So that is crazy. And I guess everything should be questioned right now because there are certainly things that are pointing to the is this a bubble question. One trend I am consistently seeing in conversations right now, especially within house teams, but also investors is that while there is a lot of excitement around AI workflows and agents, it is not reducing the need for core systems. In fact, it is reinforcing it. So people are seeing demos and still asking for a bit of a command center, where is the dashboard? Where is my view across the team? I can't just work with the chatbot or agent. So there's a real requirement, I think, particularly in legal for deterministic outcomes. You want auditability, you want clear process controls and the teams need to know where the content came from. Who touched it? Why the decisions were made and how those things have evolved over time. So even as AI gets layered in the systems like I guess CLM for legal are effectively becoming that control layer, the place where the workflows are structured, where the collaboration can happen, where the decisions can be tracked and to end. So while the AI plays a huge role in generating and accelerating the work, it does not replace the need for the governed environment. And I know every time I post something, everyone chimes in and screams the word guard rails, guard rails, with AI you need guard rails. So I guess if you think about it that way, that governed environment is still required so that you can get that version control and auditability and accountability that you are looking for. This episode I have a conversation with Winston Weinberg of Harvey. And I think we touched on this in the episode, but I noticed this beforehand. There continues to feel like a growing need for this entirely new group of professionals. I think we've talked about it before. You can call them legal engineers. You can call them legal off people. It is some segment. Call it what you will. But as pools and let's say AI offerings get more sophisticated, there's just more and more of a need for some sort of professional services or customer success or in-house ops teams that are just resources that help you get that stood up. So everyone just has been saying that they need someone to help identify use cases, understand the work well enough and understand the tech well enough to get in there in a hands-on way and to build things out. And then you've got to maintain that. So those are real resources that I think everyone needs to start thinking about. There's also increasing conversations about data and how to ensure your data is ready. Winston and I also dig into the messy reality of legal data during this episode. But who owns it? How do you clean it? Why most of it probably is not very useful unless you have some serious curation involved. If you've been listening to me for a while, you may have noticed that I talked for a long time about the need for what I believe is like a knowledge curator type role. And I would have thought by now, this would have been the most important role in this day and age. But I have yet to see any job descriptions or postings for anything like this. So maybe I'm wrong about the knowledge or the data curator or maybe I'm ahead of the times. We shall see. So other things, since our last episode, I spent time at a legal gen AI conference at Stanford. I had the chance to moderate a wonderful panel of speakers from in-house teams. I also heard from a mix of tech companies and law firms. One thing that I observed-- and this is just my take away-- there's a lot of impressive innovation happening. But it's also very uneven. So you certainly see pockets of really advanced capabilities. So we have firms rolling out polished self-service AI chat interfaces. But at the same time, in another conversation with the same people, you're going to find that these organizations are often still lacking basic things, like the ability to do collaborative document editing. So you end up with this disconnect where some of the experience-- and this is in-house and in law firms-- some of that experience is very modern. But on the other hand, some of the underlying operations are still very manual and inefficient, with just a lot of unnecessary friction. So what that tells me is that despite strong progress in isolated areas, there's still a big need for continued investment in some of the basics. OK, so there's more. So finally, this is a big couple of weeks. Anthropic did a webinar on how their legal team is using Cloud for Legal. We had a whopping 20,000 people registered to listen and learn from them. I thought it was a great session. I know teams that I taught you are having a lot of success with Cloud. The Anthropic Legal Team themselves advise you that in order to get the best value out of their tools, don't just take their plugins and their skills right out of the box, which you can. But when I agree, you still need to configure those to your own department references. And that's how you get more value out of it. So that goes back to the question where you posed earlier. Who on the team is doing that for you? We need this whole new set of people to be able to address that new body of work that is emerging. It is just the most exciting time. OK, so now, today's episode. It is a really, really special one. I was so thrilled to be able to sit down with Winston Weinberg, the CEO of Harvey, which arguably is one of the most talked about companies in Legal Tech right now, and maybe even in tech generally. And one that has truly, I think, brought this industry into the mainstream. We start with what it really feels like to be at the center of this kind of momentum and pressure. From there, we get to the big. It is Harvey, just a rapper on top of foundation models. We get that question out of the way up front. And then we spend a lot of time unpacking how law firms and in-house teams are actually adopting AI, where they're similar, where they're fundamentally different, and why the incentives, not the technology, maybe the biggest barrier to the change. We also talk about the rise of the legal engineer, which I mentioned, and this kind of entirely new career path, the tension between personalization and standardization of our tools, and why AI adoption is about so much more than buying the tools. It's truly about rethinking how the entire organization operates. From there, we zoom on to the bigger picture, and there are too many nuggets to list there. You just have to trust me that this is a wide-ranging, honest, and at times actually pretty provocative conversation about where legal is headed, and what it's going to take to get there. So I had so much fun talking to Winston. I really hope you enjoyed this conversation. And first, a quick thanks to our sponsor. This episode of Pearls On Gloves Off is powered by Workday. For years, the world's leading enterprises have relied on Workday to move their people and their money forward. Now that same momentum is coming to legal. With the acquisition of EVISORT, Workday is helping legal move forever forward by turning manual workflows into actionable intelligence, less complexity and more clarity for the modern enterprise. Visit Workday.com to learn more. And now onto my conversation with Winston. All right, Winston, good to see you. Yeah, thanks so much for having me. I have to ask you this, because I was at the Sequoia offices last week, and I was waiting for the person I was going to meet with, and on the conference room they have on the screen, like this picture that says, we helped the daring build legendary companies. And there are 12 faces on this poster thing. And you were one of them. So I had to take a photo and text it to you. But the other people-- there's only 12 people on this, right? And we've got just to name a few. Jensen, Sam Altman, Brian Chatsky, Steve Jobs, Elon Musk, Larry Page is on here. I mean, you are featured on my-- I should not be on there. [LAUGHTER] The short answer to that is, I should not be on there. How do you feel? There must be so much pressure. I mean, you have done just, first of all, incredible things at this company for us in the world of legal tech. Like you have put us all on the map. You have made it sexy and cool to be in legal tech. It is now a mainstream topic that the whole world talks about. So you have pioneered all this stuff. But that must feel like an enormous amount of pressure to be the CEO of the company that is leading in one of the most talked about companies in the one most talked about areas right now. What does that feel like? Yeah, I mean, it's definitely a lot of pressure. I'd say I'm very honest. to be even mentioned among like any of them, even if it was like somehow a mistake. - Your face is like next to Steve Jobs. Like that's for sure like some intern or something like that, like accident, like put that out there. It's quite strong, yeah. But I mean, I think like as a company scales, I think you feel a lot more pressure. I think it's less from investors to be honest. People get this pretty wrong. Investor pressure is actually strongest in the beginning of a company. - Oh yeah, I got it. - Yeah, but it's weird 'cause I think a lot of people talk about it the other way. They're like the bigger you get, like investors put more pressure on like charging more and like all these things. It's actually the complete opposite where investors put tons of pressure on the beginning and then hopefully you preview yourself a little bit and they're like, all right, they know more about the business. - Yeah, they're already celebrating, right? So yeah. - It's not even that. It's more just like they start saying, oh wow, this is a lot like a bigger company. They probably understand their employees better, their market like things like that. In the beginning, you have to like really preview yourself and you're always doing that. But my point with all of that is I feel like the pressure shifts from investors to like most of the pressure I feel is employees internally in customers. Like I feel a lot of pressure there and especially from customers in the sense of like, I think that we're gonna go through a pretty large transition. Like we've already gone through a pretty crazy three and a half years. I think the next two are gonna be even crazier and I think this year is gonna be pretty crazy. And so I think there's a lot of pressure there on and how do you make sure you go about this in the right way? And it's not just product, right? Like product is 90% of it, but I think there's a decent chunk of it that is like how do you actually work with the industry to navigate these issues? And I'd say, you know, if product is 90% and 10% right now, over time, you obviously that's gonna change to some degree, right? Where I'm not like product is always gonna be the number one thing. But my point is there's gonna be a higher percentage of your time that you're also focused on like, how does this transform the industry? Like how does this work at law schools? You know, I've been talking to do dishieries and things like that. And it's like that I think is gonna become of an important piece of this too as it goes on. Yeah, great. And I wanna get into all of that. But like the first major pressure that I feel like you're probably and I've heard you like you had to answer this answer this question every other day if not more frequently. But you know, the pressure from the question about, is it a wrapper? How does it differentiate from everything else that's out there? Let's just address it head on like why is it a tarvey wrapper? You know, on top of models that you are partnered with. So one thing to maybe start with this is like of all of the pressures, this one is the strongest but the most familiar. In other words, like Gabe and I have been shouting this from the rooftops. Press play. Literally day one. We've always said, you know, we get a lot of questions about like who are your main competitors. And I always am like, it's anthropic and open to stuff. Like 100% and being very honest about that. The only reason that helps is because you can start making like long term decisions based off of that. But the main thing I'm gonna say here is a couple different pieces. One is I think that all of these companies and not just in legal but in other verticals are gonna turn into what I would honestly refer to as closer to like infrastructure companies. In other words, they are storing all of the data. They're making sure the data is processed. I think another thing that's gonna get really difficult is the coordination between humans and agents. And for in-house, I think this is especially interesting because there's a third party which is a law firm. And so like a lot of what we're gonna build over time is take like a bunch of all of your internal data, connect that up into Harvey. And then you come up with evaluation frameworks and like routing frameworks for does this task go to an agent? And then what percentage of that does the agent do? And what percentage of that does the human do? And sometimes does that human in-house or is it an external counsel? And so like a lot of I think and by the way on the law firm it's the same but mirrored. Where what you're gonna have to do is here's all the data related to this matter and all of the similar matters. We're gonna pre-process all that work and then coordinate across a team of lawyers working on that matter. My point with this is I think that there is a very intense battle for the productivity layer. Right? And you can think of these products as like a productivity layer for the most part or a lot of them have been. And then how do you transform to an institutional layer? And in that institutional layer it's like data, workflows, evaluation frameworks are gonna be really complex and interesting, coordination, collaboration between different parties and agents. That stuff is incredibly defensible especially in a highly regulated like legal. The productivity layer is not defensible. Yeah. And so it's basically how quickly can you transform your product into that? By the way this is the same in medical, the same in like every vertical in my opinion. Think about how Eric Traffic Control works at a busy airport. Every plane taking off, every plane landing, hundreds of flights moving at once, and one control tower making sure everything stays coordinated and nothing collides. Now imagine every piece of legal work moving through your company like that, contracts being drafted, sales asking questions, vendor agreements coming in, decisions happening across the business every day. Without a command center, legal is reacting to everything. Word Smith is the command center for in-house legal. It captures every legal request, manages every workflow, and applies legal policies across the business in one platform. Right inside tools your teams already use like Slack, Notion, G Drive, and Microsoft 365. Legal sets the guardrails, the business moves within them safely and at scale. If you run an in-house legal team, it's definitely worth seeing. Just go to wordsmith.ai and book your demo today. Tell me about the agents because this is kind of a new thing for you guys, right? There was announcement last week and so you're building agents that are like a co-work type agent. Yeah, it's closer to it's much closer to co-work. The way to think about it is co-work is basically doing a bunch of things to maneuver around your desktop, right? And so like it'll edit documents and things like that. The way that I use co-work is I basically treat it as like I'm going to do a bunch of organization on my desktop and then I'm going to go into co-work. You can think of like what we're building is your desktop is Harvey. Yep, right? Yep. And then we're going to basically orchestrate over the entire platform. And eventually what we're trying to build is in the same way that you set out your desktop and all of your folders and everything like that. You're going to do this on a matter-by-matter basis. So we'll pull in all of the data that's relevant to a matter and then we'll connect long horizon agents to that, right? And it's the same within-house. It'll basically be like we'll connect to all of your internal systems and everything like that for a process and then we'll basically connect the agent to that. And so again, it goes to like something we've always talked on the product side, which is you want to expand your product surface and then you just massively collapse it, right? And so you expand it by building like every single data source that's relevant, all of the internal external data sources, all the processes, all the permissions, all of those things, different product lines to serve research versus drafting versus diligence versus litigation, and then we'll go vertical even more. And then how do you collapse everything by teaching the agent how to use the different parts of the product? Does that kind of make sense? Yeah. So who would build these agents? Is it like Harvey out of the box? Is it the client? Whether that's the law firm or the in-house team? Are they different for each? Yeah. So basically we're going to we're going to build tons ourselves. And then so we're going to build a bunch ourselves, right? And you can think of an agent as like they can access all these different skills or workflows, right? And so what we'll do is we'll build all the skills and workflows and we'll build generic ones and we'll do like practice area by practice area and then domain and use case by use case for in-house, right? And they'll be separate. And then what you can do is you can just use like the entire product Harvey will be basically accessible by long horizon agents. And so you could just type in your query and then we'll go and grab the this appropriate skills and workflows, etc. That's all generic, right? And generic, but then generic for vertical and practice area and use case. And then the law firms and in-house teams can create their own. So you want to add both because the reality is like a we can't create every single agent for every single use case. B there is no use case that is exactly the same across every single company or law firm. Similar, but it's different enough. Yeah. Which is kind of why I ask because I think it is different. It's even different amongst each lawyer within the organization like they want to totally. And so the resources that are needed to be successful like right now or at least let's say a year ago when you bought a Harvey, you could just like turn it on and give everyone a log in and people could just start working at it. But now when you're adding additional value when you're adding agents and sophistication to the product, obviously you're going to need a little bit more training, you're going to a little bit more customer, you know, professional services to get it up and running a lot more connectors, etc, etc. Both inside and within your own company. Does that also start to change? You can't just buy a thousand seats and just get going overnight in some ways you can, right? But to get the real value out of it, there does need to be a little bit more investment on both sides. 100% and this is actually you can think of what we're doing as we're building out the product complexities and then we're also figuring out how to build like services, right? And the services it won't be like we're not becoming a law firm, but what we will do is go into a law firm and say, hey, this is how we would transform your law firm. And the thing for law firms is like it's practice area by practice area, right? Like it has to be it won't be hey, how do we transform the entire firm? It'll be this is an AI native M&A practice and this is how you do it. And a lot of the, you know, consulting or transformation that we'll be doing is based off of how do you clean up all of your data? Get the correct precedent, right? Because like just connecting this to a DMS or whatever your system of record, it's like 70% of the data in there is probably like completely useless. Yes, exactly. So yeah, so a lot of it is going to be like practice area by practice area, like what are the good templates? What are the first workflows that we're seeing in the market that like bring the most ROI things like that? And then we're doing the same thing on in-house and you just have to verticalize it. You have to verticalize it to the point where it's like, "We'll have a group that does it for banks. We'll have a group that does it for tech codes. We'll have a group that does it for retail, etc." It has to be that verticalized. And so is that a product team? Is that a customer service team? It's both. It's a combination of product and legal engineers because you need the domain expertise. For anyone that's done, try to do custom software and legal. If you just have engineers, it's actually a huge problem. You need the translation layer. You do. And if you just have lawyers, it's also disaster. Also disaster. So you need to basically like marry the two. And that's a lot of what we're working on is how do you build that muscle out? We just hired a chief strategy officer who's helping a lot with that. But it's basically how do you build out a version of Palantir that's very specific for law. And a decent amount of it is not going to just be coding. It's actually going to be like change management and transformation and things like that too. Yeah. So I just like I wrote a piece because I think this is the biggest, like we talk so much about oh, yeah, it's going to like destroy jobs and like change everything and no one will have any work. I actually think there's the entire market for these legal engineers or call them what you want. I think people are debating what they call them, right? But every company, whether it's in-house, whether it's a law firm, every single legal tech company is fighting over them because we're going to need tens of thousands of them if we're going to be able to do what we're all talking about right now, which is really cool to me because for two reasons, one, it's another career path. Totally. Which is like awesome. Like that's really cool. Like there haven't been. It's incredible. Yeah, you used to come on a law school and you had one option. Now you have so. Yeah. And so I think like that's really exciting. And the other thing too is my guess and this is already the reason why I think this will happen even more and more is because it's already happening at Harvey. We have legal engineers that are just PMs now. Like they basically just transitioned into a different role. And I think that'll happen too. So I think like, A, it's really cool because we're creating this completely new category of role. And B, I think you'll see a lot more lawyers. A, join those other ones. But also B, join basically like going to be a PM, going to be on sales or something else. And I think that's really interesting and really good for the profession, especially again, like folks at a my age. I think it's incredible. Legal tech as an option wasn't, you know, there was like four companies before and now there's enormous amount and the demand is huge. There's also this role that I've been talking about for years even before AI came about that I thought was really important. I said, the most important role in the future is actually going to be the knowledge curator. And I, you know, I haven't seen that start to happen yet because to your point, right, you have this I manage your net docs really have this huge repository that has all your treasure trove of information. Yeah, 70, 80, 90% of its junk, you have drafts that are abandoned, right, that you have stuff that failed. But there's no one saying these are the nuggets. Like it's only this 10% that we should be using. Oh, I totally agree. So I think like there's the data cleanup, right, which is super important. There's the ongoing data cleanup because like another problem with this is like a lot of the problems with these systems of record or human error. Yes. It's just like you just say it like. And so if you build a new system of record or something like that doesn't change the human error. Right. And so I think that like the actual doing it once the ongoing issue is going to be a huge issue. And then the other thing too is evaluation, right? Like this is going to be super important of say you build all these AI agents. How do you evaluate them? Yeah. Right. And how do you evaluate them specifically in your company or specifically your law firm? That's going to be a huge problem and an ongoing problem. And then the last one and this is maybe two future looking and we'll see what happens here. But my guess is the legal team ends up being like the stewards of AI adoption across an entire company. And the reason I say this is right now laws in charge of basically enforcing and making sure that you comply with all the rules around human employees. There's going to be a bunch of rules about around agent usage. And like in banks, you're going to have all these audit trail rules about like making sure an agent like did XYZ and you can go and track the usage and things like that. And so I actually think that like maybe that's two future looking. We'll see what happens. But my guess is the legal orgs in top inside corporate corporates are going to actually be in charge of like monitoring AI use like making sure that we follow all the rules like all of those things, which is super interesting. This episode is brought to you by Bright Flag. Everywhere you look, a new legal AI solution is popping up promising to redefine the way your legal department operates. So how can your in-house teams sift through the noise to find the tools that will actually make a difference? At Bright Flag, AI isn't just a buzz word. It's the patented technology that's been powering their platform for over a decade, making it the most experienced legal AI tool on the market. Bright Flag's intuitive AI-powered enterprise legal management platform helps corporate legal teams manage matters, control costs, streamline invoice review, and gain complete visibility into legal work and spend, featuring hundreds of thousands of hours of training and refinement. Bright Flag's AI enables automated legal invoice review, deeper data insights, and more strategic decision making, giving your legal team the time and resources it needs to focus on the work that truly moves the business forward. To learn more about Bright Flag and book a personalized demo, visit Bright Flag dot com. Okay, so you mentioned the system of record and the audit trails and collaboration in all that is so important and maybe getting more and more important. And yet we still hear the SaaS apocalypse, SaaS is dead, AI agents are all we need now. I have an opinion on that, but do you think SaaS is dead or is it actually more important than ever in some ways? Yeah, so I think like SaaS in general is going to be much, much bigger than it is now. In 10 years, it'll be massive. Yeah. I mean, the easiest, let's just look at legal. Like the legal tech market. It's all over the place with what people say, but it's like what 24 to 30 something like that it's kind of all over the place, billion, and the legal services market is one point, you know, one to 1.1 trillion something like that. And so tech penetration really low, that's definitely going to increase like for sure. And I think that's going to be the case across like every single vertical, but the real risk I think to legacy SaaS is actually the AI native teams, how they're changing, how they ship internally. And I think it's actually the bigger risk, which is basically like most of the AI native companies that I know, especially the co gen ones and then the labs are definitely doing this. They're getting to the point where they've created different ways of shipping software. Right. And I think that's just very difficult to do if you're a legacy incumbent, right? And that might compound at some point. Like you might get to the point where like by the end of this year, AI native companies can ship 10, 20, 30, 100 times more, right? And so I think like that's where the risk is more than anything. I don't think it's like companies vibe coding internal product because the big from the vibe coding is like building it once isn't that hard. Yeah. It actually never it never has been. It never had. It's never been that hard. The problem is just like making sure it's scaled, updating it over time, investing in it, right? And so I think like that's less of a risk, but I do think there is a risk just in the sense of like anthropic is shipping something every day. That's great. I know it's crazy. That compounds, right? And so it's like that type of thing, I think, is a risk. Yeah. Okay. And yes, I totally agree with you on the vibe coding thing because we've had hackathons for like a decade and you can come up with like 20 things that are great and you know, you run a company now. You're not going to ship any of those because I'm already, but they were great ideas and you can hack them together pretty quickly. Totally. And so it's really, really, really, really, really, really important to be able to do that. And so I think that's really important. And so I think that's really important to be able to do that. Are these two totally different worlds? Are they starting to converge? Like when you're thinking about the product, when you're thinking about going to market, how are you treating them different or the same? So yeah, really good question. We have a separate product team for both now at this point. There's a lot of overlap, like a lot and a lot of overlap. I think that there will maintain to be a big product layer that is an overlap between the two. I think the more interesting thing is we do a lot of things that are collaboration between the two. Right. So basically the way to think about how our product is organized. I mean, it's actually broken into way more pods than this. But right now, the really high level would basically be there's in-house, there's law firms, and then there's connective tissue between the two. Right. And infrastructure would be one like document processing is going to be the same, all of those, the platformization, governance, et cetera. But yeah, do you think these start to look pretty different over time? I mean, the best example of this is like automating a bunch of contracting use cases is like really interesting to in-house. Right. Not very interesting to law firms. Right. There is a lot of delta between the two over time. Yeah. And you sell to both sides even though there's a connector. So I guess as you think about it, you could have like a law firm buy it and then say, "Hey, clients, the benefit of the company." of being with us is that you get access to ours or vice versa. The client could buy it and say, no, I need you working in our stuff. So I need you to dial into my Harvey. Is that playing out in either or or both both? Both. Both I'd say I think like, I mean, right now shared spaces is being used to wind deals like for law firms. So what they're doing is they're basically building a bunch of custom workflows, vaults, things like, I mean, all these are going to turn into like agents in general. Yeah. And then they're basically sharing that with a in house team and saying, hey, this is like how we did this AI native and that gets them more work, right? Whether it gets a more work on that initial pitch or gets them more work after they completed the client matters close. But so I think it's actually it's happening very fast on the law firm side doing it. We also have in house customers that are starting to be like, hey, we want all our data to reside here, right? And so we've been looking at like, okay, well, maybe we'll deploy in their VPC and then have law firms log in. It really is all over the place. Like there's some in house teams that it's like, we 100% just want the law firm to handle everything, store all our data and there's some in house teams that are like, no, we want more control over these things. So we just offer both. And what about like the adoption versus experimentation? Are you seeing different levels of like, what's the attitude and the adoption levels on both sides of the eagerness and like, yeah, I mean, you know, there's definitely different incentives, right? Obviously, interestingly, like usage is a little higher on in house, but it's not by tons. I would think so. Pretty high. They're in some huge to do that. It definitely is. I think the bigger the biggest thing that we have is in house accepts product updates faster than law firms. That's like the number one Delta, I think. But we've been doing a lot of work to basically help law firms adopt those features. That's the biggest Delta. I see right now. What do you mean adoption of product features? I'll give you like a really clear example of why this would happen. Like imagine you're turning on memory. If you're an in house team, it's like much easier to turn on memory, right? Because you have one client. Yeah. The client is just a company. I see. Whereas if you want to turn memory on at a law firm, it's like, you have to get signed off from every single client that you're going to actually use that with. So it's much harder. So there's just just by the nature of some of these features, because law firms have, you know, ethical walls, they have all these different clients. There's just some things that they are going to adopt more slowly. They adopt it really fast in a sandbox way, but it takes a little bit longer to go GA. Yeah. Okay. That makes sense. I could see that. And the shared spaces example that you gave, I've actually heard about a couple really interesting shared spaces examples that firms have put out. Why would they do that? You know, thinking about their incentive and how they sell their work, like giving access to their secret sauce or giving access to self service from their client to me is exciting. Is the right direction? Is everything that they should be doing? But I would imagine that's not easy for them given, you know, to protect some of that. Honestly, like all you need to happen. So usually it's not a firm leading this. It's like a couple partners that will lead it in a practice area, right? And then what ends up happening, because the CIOs are always like very interested in doing this. Of course. Like they're super interested in doing this. And then you get blocked by like a particular group of partners usually. And so what I've found that happens is once you get one partner on board because of a client, then all the other partners see the value in it. And then they're willing to do it, right? And I think that like firms are getting much more okay. First of all, like in shared spaces, you can basically decide what to share. So like all of your prompts, if you want, like you can hide them from the shared theme, like whatever you want. Like you can basically have all these different settings for what you actually show to your client. And so I think like those options are going to present this and you have auditability and all those things. But I do think law firm, they're starting to think, Oh, like what is actually the value that I provide? And some of this stuff is going to get commoditized for sure. So let me just be up front with my client about the commodization and show them like, Hey, trust me. Like I'm going to do I'm going to use AI for everything that I can. You're going to get the best service possible. And because of that, you're going to come to me next time, right? Yeah, you just get a couple of examples of these and then it spreads and people stop getting so anxious about it. I think that's right. And I applaud the like the handful of partners who are doing that and leading the way. Cause I think you're right. Like once you get used to, you know, okay, actually, this is great. It's actually generating more business. And it's a deeper relationship than it gets better. But there's so much like preservation of everything when people get scared. They just start to hoard work and hoard information. Oh, 100% and I think like also you get last questions when you do it. Like if you just come like a lot of what in house teams are doing right now is being like, Hey, how are you using AI? Right? And if you can actually concretely say this is how on this matter, the rest of the questions usually go away, at least for now. And so like I think like people are starting to realize that really like they just want to see actual examples of this instead of it's like kind of hard for us to measure it. It's unclear what we're using it in this case or that. And so I think that goes like a really long way just in terms of like client law firm trust. You just reminded me of just the fact that every client is asking like, how are you as a firm like measuring ROI? And when do we get a start scene in our bills? How are folks answering that question? How do you measure ROI? So I guess this goes to what we were talking about earlier where like the company, you know, you're like 90% focus on products and there's some like product transformation mix that you're going to have to start doing. This is part of like what we're building into the product is like better ways to actually track this. It's hard right now. Like it's really hard. I think that it's going to become easier when you have agents that can do a task from or a skill from start to finish because then you can actually say, Hey, this is how long it took in the past. Right now the problem is like say you're working in a word plugin. It like saves you time some like in a decent amount of times, but it's like what percentage of that was the actually the matter, right? Like it's just like really hard to quantify. But once you get end to end systems that actually do part of a task, it'll become much easier. Yeah. It's so hard to measure ROI on any of this stuff. And I think what the clients really want to know is like, how is this going to start impacting my bills? You know, you guys should be able to scale more now. You guys should be able to like do more with less and spend your time differently. And the association be turning and burning hours now. So dot like why isn't this looking different on my bills? And you know, I think that there's still the incentive like pricing is easy. You can go to fixed fees and change pricing, but the internal incentives of a law firm are still fundamentally at odds with efficiency. Oh, 100% and I think like my gut, like I don't think the bill of hours going to go away instantly, like at all, I think there will be transitions though. Like I think like people will start if you're doing like a large M and A, though be a certain percentage of all the tasks on an M and A, they'll be fixed fee. And then the rest will be bill of the hour, right? And so I think that'll happen. I think just another thing to where, you know, for a lot of legal work, like the speed and quality can be such a massive differentiator that I think will start getting ROI metrics that are tied to that. And that's going to be really hard, but I think you can get there over time. And that's where I think like getting to this, like how much emphasis are you putting as a company and as a law firm or in house on actually just evaluation. And I think for a law firm, this is going to become really interesting is like you can think of the problem as how do we connect these long horizon agents to all the data to defer, right? And then how do we evaluate basically like their ability to do work, right? And then once you have that, you compare that with, okay, let's do agent plus human versus just human. And you can start looking at like, oh, wait, actually when we have basically, you know, AI plus human lawyer, it's much faster and way better performance. Like one thing that I think is going to happen and this we're definitely not at that stage yet is to people don't talk about the fact that like legal is mostly cyclical. I mean, it's counter there's counter cyclical and like litigation and bankruptcy. But it's like if the economy goes like this and gets super complex legal work, it's this. Yeah. And so like there's a world in five, I think for sure in like five years from now, I don't think a human lawyer can analyze a data room. Like I actually don't think it's possible. Like I think it will be like think about how long the contracts and how complex they will be and how many there will be and how much business and things like that. And so I think we're going to get to the point like very quickly where it's actually like very difficult for just like humans to do this work and law and medicine and all these other areas without AI. Like I think it's going to be pretty much impossible. And we saw it in e discovery already. Right. It's just not physically possible. Yeah. You give us 10% key see instead. Right. So yeah. Yeah. Another big question I have. Let's just go back to data overall, although quality questions are really interesting. So side note, let me just my mind is going in so many directions, measuring the quality of legal work, whether it's of a human or of an agent. I don't think anyone has cracked that net yet. We've been talking about it for decades. It's an imperfect. It's an art, right? Like there's no right answer. You never get to 100% and sometimes you don't know if it's any good until years later. Oh, 10 years later. Yeah. No. So I don't think you're going to be able to do this for everything. I just think that there's going to be some use cases that you can do. So like the thing that we're starting to do and we've been doing this for a couple of months now is the coding models are getting so good that you can create really good synthetic data sets, like really good. Right. And like we don't train on any of our customer data unless we have an engagement with them to do that. Right. So we as like a default, we don't do that. And so the really hard thing unlike code, there's just like no training data like at all. And so what you can start looking at is actually you can use these coding models to create really good synthetic data sets for something like a data. Root right. And so I do think you can get to like performance metrics on like diligence like for sure. Can you get to performance metrics on like did we get by that company in the right structure or like did we. like you get that deal of, I don't know. And the same with the litigation where it's like, they sell for $500 million. Good or bad? Right. You'd have to like run a bunch of just simulated litigations and be like, this is the barbell. Right. I think you can kind of do that, but I think that's gonna be quite hard. And it's eventually always subjective. I also have this philosophy that like our AI, right? It's very good right now. It's gonna continue to get better exponentially. There's a point where good enough is good enough. Like we focus so much on like this perfection quality, but at least on the in-house, it's take this apart, right? The law firms, they do have to be close, as close to perfect as possible. That is what you pay for, that is their job. On the in-house side, you just don't have that luxury. So as many people move from being an attorney at a firm, then they go in-house, like they'll tell everyone, we'll tell you the biggest lesson they have to learn is being able to make decisions and judgments based on imperfect data. And in the time frame that you have in front of you and you gotta make a decision, you can't wait till Friday. It's like right now, what do you think? And you have to weigh the word risks and just decide. For in-house teams, at a certain point, the AI gets you that 80, 90% there, you could then turn to the law firm and pay for the extra judgment that risk transfer, that sort of like make it absolutely sure. But if you don't have the time, sometimes good enough is good enough. And I think we're gonna get to the point where like our risk tolerance and what we think quality is and how do we measure it and what does it mean? We'll start to shift because you think about like all these examples where they put like human in the loop for AI and for agents and you wanna check it and check it before the machine takes the next step and people often say like, or even the startup companies I talk to you like, they say yeah, we have to say we have a human in the loop because the clients want that, the customers want it to feel good, but like let's face it after six months, they all go it's good enough or it's so good, like I never have to check it and then they just let it go. I think that's gonna happen very quickly here. Do you agree? - Totally agree, it's literally the same as working with colleagues. Like when you start working with someone and especially they're like junior, it's just like yeah, I'm gonna check all of your work and check everything, I don't trust anything. And then eventually, I mean, hopefully not that bad. But like you do it in a little bit, like when I was a junior associate, like, I mean I'm sure of like every single property, all of that was checked at all times, right? And then at some point it's like they just trust you more and they don't check everything, right? That's I think 100% gonna happen. And a lot of what we're starting to do too within house is like what is your risk tolerance for different things? - Right. - So like NDA automation will probably happen very soon because risk tolerance, like the tech is close to being there. It's really close, weirdly it's like actually a hard problem. But it's like pretty close to being done, especially for like large asset managers and the preference is there too, right? Like you actually have to have both. Like the tech has to be there and then you have to have preference. And I think the preference or the risk tolerance, et cetera, is like fine for NDAs or and I think then the next will be like MSAs will have a certain degree where it's like you just need to check XYZ and especially if a deal is a certain size, you only have to check these things, right? And different companies are different risk tolerances but I think that will happen. - Yeah, okay, so this brings us back to like the data and the preference of the preference that you just mentioned. This is a genuine question that I have. So I love these podcasts because I get to ask my dumb questions that I'm like, how does this work? So you have all this data, let's say at a law firm or within your in house as a client who owns that data at the law firm? They have a ton. Are they just allowed to like have at it and connect Harvey to and do whatever they want? - Yeah, I mean, they basically have to get signed off. It depends on what they're like outside, you know, they have basically like guidelines, right? In all the engagement letters, yeah. Every single one of them has guidelines, right? - That's not very clear though, you know, it's like we have a data but like. - Yeah, they can be, so like maybe using AI versus who owns the data, like definitely two separate questions I think. So using AI, like I think we're getting closer to the point where clients are pretty clear about it in terms of like whether they allow it, whether they don't and on specific engagement letters. And it's across the board, like some folks are saying you have to use AI and show me how you're using it, right? So I think that's, that will get solved quite quickly just through engagement letters. I think the IP issue is an IP issue. I think it will be super interesting of what happens. One thing I would think about those in in house side is like, you do want law firms to have like some ownership or ability to use this data because you want market data. You need the market data. Like it's really important, like if you're a bank and you're entering a new territory, like new jurisdiction, it's like you'll want to know what other banks have done, right? Like that is really good. And so I do think that like, I think that this will also get worked out but it'll be messier than the first question of can you use AI or not on this and connect Harvey to it? Like that I think is pretty relatively easily solved by engagement letters. And I think to be honest, like the market has kind of settled on this. The second one is going to be back and forth massively. And I think it'll depend on different types of client matters, different types of data, who the client is, is really interesting, right? Where it's like asset managers really need private data sets because most of their deals are private. So it's like they want like to use a law firm that has all this data on it, right? So yeah, that's where I think this will go. It'll definitely get worked out, but it'll be a little bit harsher. I like the way that you use that example of AI and switch out the word AI for human. So if you think about like, I'm a client, like don't train on my data, don't train on any of my stuff. But if you put an associate at a firm, they're in the docs, they're reading all the different clients, they're getting educated, like that's how you train. They're trained on your stuff on your data. And you want them to read the competitor stuff and the other industry, like all the other stuff in the financial industry so that they are better. How is that different? Like don't you want the AI to be better by able to see? And it's not like they're going to point to, this is the way Goldman Sachs does it. And I think it will just get basically, there will be different swaths of this of how it gets decided. And I do think it'll be decided like client by client industry by industry, et cetera. And basically client matter by client matter is I think how it'll be done. But it depends. I will say like some in-house teams are rethinking their data strategy because of AI, right? Where they used to store all their documents at a law firm. Now they're thinking about, I want to store some of them. That I really want like, search applied to them, but I also want to own some of that search. It's going to be interesting. And I think it is going, it's not going to be like a super simple answer across the board. I think industry by industry, et cetera. It will matter. Yeah. OK. Another question I have is about personalization and standardization across the organization or within a team. So at a law firm, I've talked to folks who are like, OK, I want my associate to use a model that's trained on just my stuff and Joe's stuff as a partner, because we're like the best. And we do things the same way. We do the same way. And so I want it to like, even like, let's again, take the AI out. For my associate, I want you to train on like our stuff. I want you to work with, listen to me and listen to him. On the in-house side, I've had like way back when AI first came out. I remember I was at Google and someone was like, wait, so like we can have these machines that can basically like negotiate just like me. And I was like, well, like us. And he's like, well, why couldn't I just make mine like just personalize to me? And I'm like, oh my god, that's what we have right now. And this is the problem, right? This proliferation of working in silos, like we're trying to create some standards. How are organizations or how do you guys think about that? Because it gets better and better if it can work like me. But it's also not what we want. This is exactly what happens. And I think like everyone goes through this. And then they end up actually coming to a conclusion you just had, which is like, you know, we need to do this as an organization. This is very similar to the problem that I think is going on like in AI across the board right now, where everyone is thinking of this as individual productivity. That is not going to scale. You need institutional productivity. And like a lot of what we're trying to do internally is like just giving everyone coding models and making sure they're token maxing to the moon is helpful. But it's like the reality is actually that's not that helpful. And you're not going to get like crazy gains from AI unless it's you change how your team works. And various teams and then how your company operates as a whole, right? And I think like that is something that we are going, it's going to we're going to all struggle with this. Yeah. But it's really important because we need to figure out how to basically do multi player an entire organization, entire team scaling with AI, not just individual productivity. Yeah. And this is a really important point, I think, like as getting out of soap box, there's a lot of great tools out there. The tools are not like the magic answer you can't just buy a tool, whether it's an old-fashioned sass or like a new, you know, AI Harvey like thing, you need to invest internally, you need to invest with, you know, your team and like, there's time at investment, there's people investment that needs to happen. Otherwise, it's not going to work. And again, you can buy tech and it'll just sit there and not be helpful. I mean, it's why like everyone needs to start thinking about this is like R&D. And I think this is like really hard sometimes for folks to do of, we're going to pause stuff and slow down all the things that are coming at us and make it so that we're going to figure out how to automate this. So it saves us a bunch of time in six months. And it's just hard. Like it's hard for companies to do that. And it's really hard for lawyers because we get inundated with so much work, like so much, right? And, you know, I actually think that's one advantage of having a legal background and then being a founder is like, it's similar in the sense of like, you have to triage, just some good billion things coming at you. But like you have to learn how to do R&D. And I think all of these different orgs are going to have to do that. And that's going to be hard. What do they need to be hiring for? I mean, it's partially like a legal operations type function, right? And legal engineers underneath that and project managers and change management. Yeah. It's like, I think like three pillars, domain expertise, right? And then technical expertise and then people expertise. Like, change management expertise, right? It's like a cross. It's like those are the three pillars that I see. If you can find unicorns that can do all three, great. That's usually hard to find. It's like better probably to find like people that can do one, like two of the three, something like that. That's what I found. It's hard to do three. - Yeah, totally agree. Although I will say the unicorns are to the beginning of our conversation, like more, they're out there now because there are people who are lawyers to have like veered outside of law just because of what's available to them now. - There will be way more for them. - Way more. - They can do all three of the pillars in a couple of years. - Yeah, I'm just saying we've created a market. - I've never done as many. - Yeah, exactly. - Yeah, exactly. - Yeah, exactly. Okay, so that makes it, I think, particularly hard. And I'm gonna do a talk for the scale up GC conference in London in a couple of weeks. And it's so funny. I've been trying to think about what to talk about in this talk. And this world is changing so fast that every time I like make a draft, I'm like, okay, it's out of date. Like things are changing too fast. But what they have sort of thought about maybe in an observation is that large organizations are going to benefit a ton from all this great technology. But it is also more difficult for them to adopt it and to make it successful because of all the stuff. Like the legacy and the change management and where all their stuff is. Versus like a scale up GC who is at a brand new company and doesn't have like 20 years of documents and doesn't have old ways of working, they actually can get away with vibe coding too because they can ship it to themselves, right? And we're like the two people that they work with. So they're actually at an advantage to adopt tech and to change way fast. I think they're gonna lead the way and teach. Like it used to be big companies. We're teaching small companies like you've learned from them. And now I think the small companies are gonna show us the way. - Oh, 100% and I think like those smaller law firms too, like solo practitioners that have done a bunch of things and things like that. Yeah, I think that's 100% right. Like I think you're gonna see like very interesting things come out of those folks. And you see it all over Twitter and things like that too. And I think like one thing that the bigger institutions could do is like either acquiring them. I'm dead serious, right? Or just like add them to their ranks in some way shape or form because I think that finding people that have worked in an AI native way is insanely valuable for every company. And figuring out how to put them into places where they can influence your company is really important. - Yeah. - Okay, so then let's talk about these AI native law firms because I think there's multiple things happening, right? There's the ones that are just starting with technology and then adding services on top of that. There are firms now we're gonna have on the podcast and I think soon that are just partners and technology. There's no associate layer. Then there's like a big name partners at huge firms that are spinning out and starting their own thing and have like MSOs that are maybe AI native. AI native is started. I mean, who even knows what that means? Anymore, but like it's starting to be really important. One question I have and I'd love your comments on all this but one question I have is a lot of them seem to have their own like they're developing their own technology stacks and AI versus there are like thousands and thousands of companies right now that are offering legal technology. You know, why, I mean, it's a question for them but like why are you developing your own versus using some of this stuff and then just adding the people on top of it? - Yeah, I mean, I think it depends on 'cause there are like varied, there's tons of different types - And some are just using Harvey. Like I know that from the podcast. - Yeah, exactly. So I think it just depends. I'll say that like we only offer certain levels of like ability to customize, right? And so I think like part of it is there is a gap in the market for just like insane level of customization like everything's in API basically. We're gonna offer like more and more of those options but I think that's actually probably the main reason. Right? And then I think for some of them like they wanna just own their entire tech stack like as much as possible. And when the problem with that is it's gonna be, it's again easy to do in the beginning. I think it's difficult to do it use scale but the more you're a tech company, the more you can do that. Although like we don't build our entire tech stack internally either. (laughing) Like we still use the clouds and things like that. So yeah, I don't know. And then my thoughts on it in general, like I think a lot of them will be super successful and they'll see what happens. Though like very simple answer I always have to this is, I think there's two types of like professional services or legal work. There is basically like I want advice and then there is complete work for me, right? And then there's a whole spectrum, right? Yep. This seems like, like this is never gonna get automated because it is the nature of advice. Like it is like it'll be easier to get advice. You can use data better things like that but the reality is like you are going to something or someone for advice, right? And then there is the, just I want the work to be done. And the question is just like how much of the work that the Neo Labs are going after is over here versus like closer to over here. I can't say it broadly because they're all doing different things for Neo firms, not Neo Labs, although there's a lot of Neo Labs too. And the question is like if it's over here you're just gonna use software. Like there's just no reason to have, like you're just gonna use it internally, right? So like if NDAs, like you're probably just gonna, if software gets there it's like why use a Neo-Strength. That little judgment needed anymore, right? - There's gonna automate that, right? And so the question is basically just like on what continuum is this in my mind? - Yeah, so yeah, and I think what we start to buy from external legal service providers starts to change because historically right, execution and let's just call it, let's just call it execution and judgment with judgment means relationships, horizon scanning, industry like all the stuff that you need the people for. - Private data, market data. - Exactly, you still need that but there's a huge execution layer that is gonna get commoditized if it's not already happening like very quickly. And so whether you're gonna do this execution in-house or it gets done by a law firm at a low cost because it is just a commodity, like that I think is happening now. And then the judgment, relationships, et cetera stuff is what you end up really valuing. And you should put a lot of value on that, like more than what you charge by the hour. Like that's a lot of value, but that's what you're actually looking to an external provider for now. - Totally agree. And I think like there's an argument to be made that a lot of the problem actually is just like how people, how law firms price things. In other words, like, I'm not like if I pay a million dollars for an M&A or whatever, I mean, actually that like doesn't bother me that much. Like that price or whatever doesn't bother me that much. What usually bothers me is like how it's been built where it's like, you know, there's junior associates going through a deal, you know, the illegitimate room, things like that, data room, doing a bunch of diligence or whatever they're doing. And that's like 15 under an hour. And then I got on a call like twice with a partner and that's like 3000 an hour or whatever. The call at the partner is actually like, I would have paid like a couple hundred grand maybe. - Exactly. - I don't know if it actually changed the transaction and it made it so it went through. Yeah. Right. And so it's this weird like pricing mismatch. - Yeah. Well, and that like that's to me why it's so frustrating 'cause the pricing piece, we can have this conversation with a lot of people and they would agree with that. Like it's obvious my GC at Google used to say, I will pay you a million dollars per hour. If that phone call makes this dumb. - Yeah, if this phone call makes this litigation go away. Like call the right people, make it go away. And yeah, it's totally value not not the inputs that I want to pay for. And the firm is like, okay, I get that we can price that way but then internally the structure of your incentive is the way that you pay your partners, the way that you pay your associates doesn't work with that pricing. So like you can change one side and not the other but then they're not incentivized to use the new pricing because it doesn't help their personal compensation. So there's this like whole flywheel of things that need to be undone internally which again is why I think the Neo firms have a huge advantage 'cause they don't have to think about how are we gonna undo this over the next 10 years? - Yeah, I totally agree with that. Although I will say on the law firm side, like it should feel good that people value it that highly. Right. Like if you think about it, it's like if you are one of these large law firms, the thing that I would go back to about feeling like safer as long as I can make some of these changes is they did value that M and A at a million. Like they actually didn't devalue that M and A, right? Like, hey, I haven't made it to devalue that M and A. It's just devalued what you're charging me for in terms of that M and A. And so there is like, I think an option here for them to do as long as they move forward with that being like the key, the like North Star, which is like, actually the value is still there. It's just we have not priced it in a way that makes sense. - That's correct, that's correct. But it's still, it's really hard for them to get there because they can't just change the pricing. They have to change the whole thing together and it's just, it's a lot to undo. Yeah, I agree with you. They should see that there's positive. - Yeah, oh, I understand. - Go have this conversation, guys, figure it out. Okay, the future of legal work. So I mentioned there's a firm that is just partners in technology with all these neo firms, like I've been having all these conversations. They're actually a partners that are at big law. You know, the ones that are spinning out, they're going to their leadership teams and saying, "Hey, I feel like I can charge the same, get the same work done." And I just probably need like one of my associates and then a bunch of good technology. So can I like price and work this different way? And big law is like, no, we need you to keep all these other associates busy. And they're like, okay, never mind. I'm just going to go pick a few partners and go start my own firm. I also have a friend that I think I've talked about on this podcast who's a big leader in investment banking and he's like the only reason we have associates here now is so we have future partners. And I think you could say the same thing, right, for law firms, like we all need future partners. We all see willing to pay like hundreds of thousands of dollars for their judgment and their value. - Yeah. But how do we get them there if we have no need for the associates anymore? Yeah, this I think is like probably the number one thing actually that I think firms need. I think this is more important than actually. The only thing that they've changed. And to me it's like, I mean, I'll do the law firm first. The thing that I always tell folks and we have a lot of law firms, like customers that are starting to really actually think about this and really change how they operate. Yeah, they are. And it just depends on who the leaders are, right? In charge of this. And the best way that I've tried to explain it to folks is like, you need to reduce your time to partner. It's like what you're trying to do. And the first piece of this is actually figuring out what makes a partner. Like what actually, and it's practice area by practice area because it's different, right? I'm like, they're sure there's commonalities and things like that, but it is like practice area dependent. And I think that firms haven't spent enough time on the apprenticeship model. Like I think like from what I have heard, like I wasn't practicing law, you know, 30 years ago, 40 years ago, but from what I've heard, the apprenticeship model was like more alive back then than it is now by a lot. And I think like that's the main thing that firms need to think about. And I mean, the nice thing is I will say that technology is going to enable this, like to a massive degree. In the sense of like, if you do have, you know, agents that can do a task from start to finish, there's no reason you can't turn that into a training module, right? And so it's how can you use technology? How can you step one is identify what does it take to be a partner, right? Step two is like of those different tasks, right? How many times do you actually have to do them? How many times do you need to be in a data room to be an M and A partner? It's not one. It's not a thousand, right? Like it definitely isn't a thousand. And so it's somewhere in between there. And then step three is how do we use technology to actually give them the learning experiences that they need? And then number four is, and this is going to be the hardest by far for law firms. Put your junior folks in high stakes environments. That is going to be the hardest thing. That is like the thing that, and this is the one that's like closest probably to like my heart just because my age and like all of my friends are mostly like associate to big while. And it's like, gives junior folks opportunities. Like they are terrified to do this, right? I mean, to put them on client calls to bring them to a trial, like to put them into a boardroom during a, like, it is terrifying. And that is something we are going to have to do as an industry. And that's something that like jumping into tech has been a crazy wake up call to me because tech does an incredible job. They let people fail. And then they learn way faster than normal way faster than normal. And I think that's going to be the hardest thing by. Yeah. So what I thought you were going to say is the hardest thing is actually getting the partners to invest their time in these junior folks because I do believe, I mean, because I've seen some of them who actually do pick one as a apprentice. And then they do become like the youngest partner at the firm because they are like able to get where they need to go faster. But not enough will invest the time because there's this, I mean, and it's not like they're bad people. It's just what's in it for me. There's like an army of you. I have a different one of you every time I, you know, have a new deal or a different litigation case. So I'm not going to invest my time because yeah, I'm not going to work with you again. I'm going to get the next guy. So there's we've got to change a lot of it. But part of it is like, yeah, the pyramid structure changes. So maybe it's a one to one or a one to two thing where you have someone that you are invested in as a partner and you care because you're going to keep working with them and you better get them good enough to like take on some of your stuff. And I think also like you could change how you do work assignments like you to make it so it's more of a pod-based system. Look, there's so many different ways that you can do this. I think that the thing that I'm looking for right now is just our is leadership that law firms thinking about it. Like that's just like number one. And I will say like over the past six months, it's been a huge increase. Yeah, I have to double do triple the amount that are thinking about it. I would say that more than the business transformation, this is a more important problem to solve by law. Yeah, it's everything. It's what the value is at the end of the day. That is worth the values. It's right. You better start building more of that. So well, this has been super fun. Winston, is there anything else that we need to talk about that we haven't already touched on? I think one over a bunch. Yeah, we're kind of. This was so fun. I'm so glad we got the chance to do this in a longer form. And yeah, I hope you'll come back as the world is changing so fast that I want to ask you 1000 questions every time tomorrow, maybe tomorrow. It's crazy. All right. Thank you so much. All right, everyone. I don't know about you, but I love that conversation with Winston every time I talk to him. I feel like we need at least an hour. So I'm glad we got to do that live with you all. The world is changing so fast. I have so many questions and I hope I addressed some of the ones that you had in your head for what's going on with the world of AI and legal and Harvey. So if you like this episode and others continue to follow us wherever you get your podcast, give us a like, give us a follow and don't forget I also have a sub stack where I like to put out some stuff and we are always grateful for our wonderful sponsors. If you're interesting in being a sponsor, please feel free to reach out and we'll catch you all next time.