Episode Transcript
[00:00:00] Speaker A: Welcome to the Relay, the show for plaintiff law firm owners, presented by lexemica. I'm Gabriel Stertz and this is my first episode with two guests. So bear with me. If I don't know where to look, it's going to be a Fun1. At Revcon, Dan Morgan announced that Morgan and Morgan is on Lexemica. Morgan is also a Foundation AI customer. Vivek and I are both sitting across the table from a guy who we answer to, so we'll try to behave ourselves here. In September, Morgan and Morgan pulled the curtain back on MX2, an AI platform that they've been running inside the firm for years. And they've committed at least a billion dollars to AI and technology over the next Dec.
John Morgan's line is that the firms that win won't just license an AI tool. So to talk about this today and other things, I have Yath here, the Chief Transformation Officer at Morgan and Morgan and Vivek on my left, co founder and CEO of Foundation AI. And we're gonna have a really interesting discussion today. Thanks for being on the show, guys.
[00:00:49] Speaker B: Thanks for having us.
[00:00:50] Speaker C: Gabriel, great to be here. Yeah.
[00:00:52] Speaker A: So we'll spend a couple minutes on this. People who've watched podcasts, I'm sure they've heard your origin story. If you're shoving ads at them, they've also heard your origin story.
And yeah, you've been a little bit more behind the scenes and inside of Morgan, but you've done some media appearances. Let's just start with your background, how you got to where you are, what your role is in Morgan.
[00:01:11] Speaker C: You know, it's actually a really simple story. I was pretty excited about the opportunity eight years ago. I didn't think I would be here eight years later.
The growth has been an amazing thing to see, but I've been at Morgan for eight years, but I've had 16 different jobs because the firm has been almost 16 different companies. Because every six months or so we grow in a very specific way. Everything breaks, we rebuild, we add new things, we take on new initiatives. It's been an amazing ride.
[00:01:37] Speaker A: Vivek, how about yourself?
[00:01:39] Speaker B: Yeah, just quick and dirty. Was a lawyer and I saw a lot of problems with high volume litigation back then in 2018. So a long time in the business as well. Problems mostly with physical mail, but also just getting documents the right places. And as we were getting more technology, it became very clear that these analog processes weren't working.
And so started this business in 2019 and work with a couple hundreds of law firms, including Morgan, I do have
[00:02:03] Speaker A: to ask the question. 2019 was pre GPT.
[00:02:05] Speaker B: Yeah.
[00:02:06] Speaker A: You called foundation AI. Were you just making it up back then?
[00:02:08] Speaker B: No. So that's the funny thing. I think if we started the business today, we might not call it foundation AI, because when people think of AI now they're thinking of chad, GPT and generative AI. But AI has actually been around for a long time. The definition of AI technically is computer systems that do work that ordinarily require, like human thought or something like that. Something close to that bastardized it a bit. But that means that by definition, the definition of AI is consistently changing because at one point a calculator like math required human thought. A calculator to be AI, or spam filter is a great example. And spam filters are similar because they are a form of natural language processing and classification, which is how what we do started in a lot of the models we use. If you think about it back in the day, and I'm going to date myself, when I first got email, like, you had to read every email and figure out if that, like, Nigerian prince, like, really wanted to send you money and stuff. But conceptually now that is AI, if you think about it. But not anymore because it's so ubiquitous that we don't think of it that way. So when we started the business, we were building our own natural language processing models, computer vision models, and we did a lot of that stuff. It started with research data scientists and I. Luckily they were the ones who built the technology and stuff. Yeah. So we were doing AI, but what people think of as AI has now changed a bit.
[00:03:20] Speaker A: That's a great lead in to back to you. Morgan just publicly committed to putting a huge amount of resources into AI. That's a very broad way of talking about it. What does that mean practically for Morgan? And Morgan, as you're investing in the future and building out, does that mean that all your lawyers will be gone in two years? Does that mean you're never going to license another piece of technology? What?
[00:03:43] Speaker C: Yeah, I think there's. I think about it in multiple ways. I think one, this is my favorite point to make. We're only hiring more attorneys, we're only hiring more paralegals, we're only hiring more case managers. If you look at our jobs board, we're only posting more and more. So the fact that we're using more AI and investing more in AI doesn't necessarily contradict the fact that we need more people. We're a growing business and actually one of the biggest constraints that we have is actually hiring.
So one of the value propositions of AI is it allows us to provide our current employees with more capabilities that allow them to do more while we look to hire more people to continue to grow at the rate that we've been growing.
[00:04:24] Speaker A: Was there an inflection point for you where you said this is we have to go in this direction where AI became useful enough that it wasn't just very limited use cases. And you said this is how the future is going to be built.
[00:04:38] Speaker C: There's a lot of problems today that we think that AI will solve. We were looking to solve those problems just with different tools maybe seven, eight years ago. So in terms of what is a business problem that we're trying to solve, we're trying to make sure that our attorneys can do more.
How do we generate more capacity for them so they can handle more cases or resolve cases more quickly?
And the original way that we were doing it is actually creating, like, centralized services. Could we create a record retrieval team? Could we create a police report team? Could we add some additional functionalities within our organization in order to get more coverage?
[00:05:13] Speaker B: Free people up.
[00:05:13] Speaker C: Exactly. Free people up. Because we don't want an attorney following up on records. We don't want a paralegal following up on records. We think specialization is the pathway there. So really, everything that we're trying to tackle with AI today in some way or the other, we were trying to do before.
And essentially it was very clear to me, because I was managing the technology org and the operations Org, that actually technology is a better solution. It just would not keep up in terms of you couldn't build it as quickly as you need to as you could with humans, at least initially.
So I don't think that this is going back to your question. I don't think this is one or the other. It's really about adding more leverage to our organization.
[00:05:53] Speaker B: I think that's. We talked to obviously hundreds of customers that are much smaller than Morgan. And it's the same attitude back in the early days of our business, like in workers compensation, when people had these large departments they were asking about, I want to reduce headcount. I never hear that anymore. Now the question's always, how many of you think it's easy to find good people? And no one raises their hand. So if you have someone that shows up every day, that pays attention, that has attention to detail to do this type of work, like data entry type work, like, couldn't they do something more high value, ultimately driving more leverage for people I think has trickled downstream where it's drive more leverage for lawyers. Then it's like, wait, paralegals can actually do way more. If they're not doing administrative work, then wait a second, we have these clerks that actually can do what paralegals used to do, follow up on records and then, and the ball kind of keeps moving ahead with what you can automate.
[00:06:38] Speaker C: The real truth is expertise matters. And within any of these organizations, it's not just attorneys who matter, it's the paralegals that matter and the case managers matter. And you know who will tell you that? The attorneys.
No one wants a better paralegal or case manager or someone to work with them than an attorney does. We think about it and kind of jumping on Vivek's point is like adding support to all of your employees because it's not just important to get more value out of your attorney in terms of providing them with support, allowing them to do more, but it's also important to do the same thing with your support staff. Because a lot of these support staffs at smaller firms, mid sized firms, are long tenured. They in fact there's so much institutional knowledge that lives in their brain and there's certain things that only they can do. So what about taking away the things that anyone can do and making sure it's done consistently over and over again?
[00:07:28] Speaker B: That's two points that I see a lot too is two things. One that you talk about a lot and so do I, which is you looked at problems that existed and then tried to find solutions for them. And then when technology got better at solving it than other solutions, you replaced people, a lot of people, with technology and so on. And that's how I think of everyone should be looking at their own businesses. Your problems that you're solving are not the same as a 20 attorney firm. Right. That's not the same as a five attorney firm. So every firm needs to go look at their own operations. And you say this a lot and figure out like what actually happens, how and to me, where our business starts when a document comes in. Because in legal cases everything is driven by something that's coming out of a document. All these inflection points that move a case forward, policy limits come in, it goes this way or that way, medicals come in. The objective findings are not everything. Moving it forward is coming in through documents.
[00:08:14] Speaker C: I think your train of thought really is, hey, by understanding the way that people work, you can actually solve big business problems. And the reality is in those documents is really important. Information.
But you want to know consistently and liably that someone reads it, puts the right thing in the right place, so it can trigger the right action. And then by the way, you design through all, all the right actions in a way that you meet your business goals. Yeah.
[00:08:37] Speaker A: So one last question for you, Yath, related to this announcement.
This is a, this is an investment announcement. Do you view this as we're going to spend all of this building our own?
[00:08:48] Speaker C: Yeah.
[00:08:48] Speaker A: Or is there a. This is. We're investing this. And what is the build versus buy question? How did you answer the billion dollars of the billion dollars?
[00:08:56] Speaker C: So first I'll say we've been investing already, but in terms of being more aggressive with our technology investment, we actually picked that up in 2022 when we saw the opportunity, specifically with AI and we were trying to decide if we were going to build or buy. And I think long term, we decided to build because of a few reasons. The first reason is we saw our needs as pretty unique, specific to the offerings in the marketplace. So that was just, hey, we tend to litigate a more significant percentage of our cases. We don't resolve in many cases pre litigation. We have a lot more trial attorneys. They have very specific workflows in litigation. We're very litigation conscious. So at that time, not many people were doing this right. So a lot of the focus on AI, specifically in the consumer law space was actually in pre litigation. Whereas we see ourselves as a litigation firm, and we, of course, we do pre litigation, but pre litigation is to set the case up for litigation when it's appropriate.
So, one, there was a lack of offerings to support our critical need. Two, we had expertise that suggests that just because of our size and scale in terms of the number of cases that we're working with, the number of phone calls, the number of documents, that it just was going to be, that our needs were going to be different structurally than other businesses as well, at least in this space. And the third was we have the benefit of thinking really long term.
It's something like John Morgan preaches. So when we looked out 10, 20 years ago and we see the way the world is going, isn't it better for us to own and manage and be good stewards of our own data and leverage that to continue to grow our business and deliver value for our clients?
And those are the key things that really drove us there. But that doesn't preclude us from buying different tools. We explore all the latest and greatest because we have such a diversified business where we're working with technology providers that support us from the call center to getting our clients paid at the end of the case. So that is just a huge work stream. We're not going to build all those solutions ourselves. And even when we do build our own things, it's not like we're building them from scratch. We use Amazon, we use other services and of course they serve as the rails to what we end up building ourselves internally to service ourselves.
[00:11:28] Speaker A: I appreciate the take very much. Do you think that it means that as you move forward that the buy side is going to be more on the infrastructure and the more agnostic workflows and the build will be related more to your proprietary data and what's happening in internally for you? John Smith, your managing partner on personal injury, yesterday at Revcon, he was talking about how the more core and bespoke the work stream is, the more you're likely to build it internally versus.
[00:11:57] Speaker C: Yeah, yeah, 100%. So think about this. We do everything from personal injury, auto accidents to Social Security disability to employment and labor to veterans benefits. So we support our clients in a incredibly wide range of claims that they may have. And obviously we're not going to be able to do it all. But our primary focus is what we consider core and differentiated. So if it's core to us, like our personal injury department, where John's been building that business for 40 years, we're going to try to service it ourselves because there's unique insights that we have, there's specific ways that we want to deliver service because we think ourselves as, you know, a unique brand with a unique product, which is a service product. So we want those touches to be very Morgan and Morgan esque.
So those we're going to try to build in house very specifically for that. In other areas, we still have a very specific point of view, but focus is a superpower. Maybe for Social Security disability or veterans benefits, there are specialized tools with people who are really deep in those workflows that have been able to spend more time on it than we have, that are going to be able to support us better than we're going to be able to build it.
[00:13:06] Speaker A: Vivek, you have customers, Morgan and Morgan, all the way down. Do you think that point of view is the right one for everyone?
And how should people be taking that perspective of there are parts of the business that we have to really be able. More specifically, how do you apply that if you're a 20 person firm?
[00:13:25] Speaker B: Do you want my real answer?
[00:13:26] Speaker A: Yes, absolutely.
[00:13:27] Speaker B: Look, I think a lot of firms, they know what they know because they know what they've seen, but the universe of what they've seen is quite small. And so we talk to firms all the time that are like, we do it this way.
And we're like, that's awesome. And it makes no sense.
This is how we do it. And how you do it is stupid. There's so many more efficient ways to do this. Let's see. And then. And invariably, like, once you. And it's usually like, once you get to someone who has, like, who actually is overseeing the whole process, usually people in insular, like, parts of the, like, fiefdoms within a business that have their way of doing things. And when you dive down, it's usually, we do it this way. And do you know why?
Because someone told me, do you know? And then you trace it back to, like an office manager from 12 years ago that's no longer there, that trained everyone a certain way. So I think when you get into, like, battle scars.
But it's so Morgan obviously has proprietary workflows. They have so much data, so much experience they put together. And a lot of firms do with maybe how they do a depot or handle a trial. The stuff Bob Simon talks about. Yeah, but a lot of the operation stuff. Firms don't know how to run a lot of their internal operations that well because they've never seen it. And also, as technology comes in, you can refactor the entire way workflows work in a way that you couldn't really understand unless you've seen it before. Unless you understand that. Oh, with technology as an enabler now, someone doesn't have to look at a document to figure out what it is before it gets routed to the right person. It can just automatically go where it needs to go. Oh, when it comes in, all the data can be extracted at once and put in all the right places. It doesn't have to go through two steps and all. So I think that certainly, like, litigation workflow and like, thought, but not every firm needs to do it their way. Some should embrace best practices.
[00:15:09] Speaker A: That's a perfect question. Or segue into the next question, which is Yath, when you launched MX2, you said you studied what separates great results from average ones. And I would say, corollary, you said, what separates great firms from average ones? How did you do that? What was that process? Did you go talk to all the law firms? Did you talk to Vivex who study law firms and look at their operations? Or did you go completely outside of law to look at how do you Build this on first principles.
[00:15:37] Speaker C: I think it's interesting because when we built, when we started building MX2, it was a very specific focus on our business, our needs.
So we really just went bottom up in terms of, okay, what are the use cases that people really need support on?
How can we help them?
And we had a couple false starts. And DJ was talking about this yesterday.
DJs our head of product. He did a great session at revcon yesterday, and he was talking about this specific product that we called Responder to help attorneys respond to discovery.
And essentially it took us three tries to get us to where we're at now. We've, of course, rebranded it because I can't stand being associated with the failure of the name Responder. I'm like, I will never call anything Responder again, although I own Respond AI the domain name.
But the.
So I think one, we had a couple false starts. And that's the benefit of working in Morgan, where we have a culture of taking risk and trying things. And if it doesn't work, then we'll try it again. Our first version was like, really leveraging very early API access from OpenAI, building UI within Salesforce to help people respond to discovery. And we learned. We learned a lot of things. Not enough real estate, because it's in a very specific component within Litify Salesforce. Then we tried a different version where it was more of a web app.
Then we found out, actually, paralegals don't want to go into anything like that because really they want this to be where they do their work, which is Microsoft Word. And in order to really get adoption, you got to move it into a space that they really want to be in, because otherwise, even if there's little things like formatting and the shortcuts are slightly different, it just adds different points of friction from an adoption perspective. So really, I think we just tried based off instincts that we have on where we can deliver value. And then when we looked and measured and tried to understand why things weren't working, we just changed course. We tried not to be so too proud.
[00:17:36] Speaker A: Let me ask you a challenging question here, which is that Yath is building for Morgan and Morgan, what happens. What has happened to you in the past when you have tried to build around a single anchor customer? And then what are the failure modes that come out of that?
[00:17:49] Speaker B: And the difference is what Yat just talked about. And I know Morgan's culture is like, when they're building, they're not trying it necessarily in production, out in the wild. They're very careful with how they build and test things. And then he talked a lot about the problems of adoption because Morgan has a thousand people that they have to get adoption. So they essentially, inside their own org, get to do more product testing than most startups get to do with something before they roll it out. So building for their org, it makes sense that afterwards they've got enough user data and stuff like that to actually have a product.
[00:18:18] Speaker A: But my challenge to you is, do they actually. Because it is different when you can impose it on your employees than when you have when you're going to another.
[00:18:26] Speaker C: But imposing doesn't work. Okay, yeah, imposing doesn't work. It doesn't work to say you have to use this product because people will find ways around it. I'll tell you this early story. I went to the call center and I was watching our intake agents work and we had built Salesforce flows and we had built a little note thing within the Salesforce flow. So if they have needs, things that come up in conversation, they can put it in this little notepad on the side. Very simple. It's like the UX is clear, you're going through this flow, you're talking to someone, you're entering data and you're like, okay, any ancillary stuff put here. And of course we wanted to track that because we wanted kind of some free text understanding of things that weren't being put in structured data areas.
No one was using it. People would find a way to use a different notepad that was better. They would write it on their desk.
The reality is if you don't have something that's good, people will actually spend a ton of energy working around your product that is not good and going to something that's better. So it doesn't matter if you're a Morgan or at another organization.
[00:19:30] Speaker B: But I think that's the point is that if you're at a smaller firm, you can compel use and then the data you get from people is not real data to the point at a Morgan you actually get real data because people have can, or at least at that point could work around things and stuff like that. But I think your question around building for a specific firm, like, it's, I think a lot of like challenges that every startup goes through. And even Morgan talked about how they do something, it breaks, they rebuild it, but they have the capacity and the ability to be able to do that like inside their sandboxes and with a large engineering team, bigger than many large startups to do that kind of stuff. Firms out There that are building stuff for themselves, though I think it's a whole different ball game, especially because many of them are not actually building their.
[00:20:15] Speaker C: Can I tell you a difference at Morgan too is our attorneys and our support staff are incredibly engaged. Like, I get so many ideas and comments from our legal teams. So if something's not good, they'll also tell me.
They're very happy to tell me. So, like, that also helps. Right. Whereas, like, you just have direct feedback and we have a culture of, hey, just go to the person directly. If you're not. If something's not working for you, no one will hesitate to let us know.
[00:20:45] Speaker B: I think one of the problems that a lot of legal. I'm sure everyone has seen this, but a lot of legal techs have is you build something for a customer and then you run into what I talked about earlier. Everyone has different operations, everyone does things differently. And so a lot of startups came out in the legal tech boom recently that didn't necessarily come from working in law firms like you and I did and, and start a company and go, great, we built it for this law firm, they love it. And then they take it to another firm and they go, none of this works. They don't do anything the same way. And none of this plugs in. We hard coded all of this.
[00:21:12] Speaker C: So.
[00:21:12] Speaker B: And then they have to rebuild. And it's, it's a different animal because firms do operate.
[00:21:16] Speaker C: Very true. There's, and I think you and I have talked about this, firms have different strategies. Yeah. There are firms very focused on pre litigation. There are firms that just handle litigation cases from referrals. So there's strategic differences. There's no one size fits all, so you have to service them differently. And also, even at Morgan, there's big state by state differences. Yeah.
New York and Florida operate very differently and California operates differently. So even at a big firm like Morgan, incredible variety in the way that things need to be done because of different laws.
[00:21:50] Speaker B: Yeah.
[00:21:51] Speaker C: And different environments.
So there is like, I'm not saying you have to have a one size fits all approach, but I think being able to play in our sandbox and having a very specific culture on building, generating feedback, iterating, taking risks. I think that does allow us to do very specific things that other people can't.
[00:22:13] Speaker A: Yeah. And there's also a philosophical piece for me here, which maybe I don't know if you and I have talked about this before, but a lot of legal tech I've seen over the past has really been built around really ultimate flexibility. And customization. The CRMs that came to market 2016 through 2018, filovine, smart advocate, Litify, functionally were infinitely customizable relative to other software that I've seen.
[00:22:38] Speaker B: Filevine and Litify, they have very.
[00:22:39] Speaker A: Yeah, they have a.
[00:22:39] Speaker B: They have no common data scheme of custom fields ever ways to build.
[00:22:44] Speaker A: I actually think that lack of opinionation in the product cuts both ways. It's easy to sell that kind of a product because you can fit Jimmy's back office and then Joe's back office and then whoever. But I'm curious to hear from your perspective where you've seen that failure point and where you think that's a product edge. Because I would assume that MX2 is more opinionated of a product than one that's just coming from, let's say a 19 year old who's coming out of Y Combinator who doesn't understand anything about law and is just going to go build to the really the requirements of every customer that he encounters along the way.
[00:23:16] Speaker B: Yeah, and I think there might be a difference between a lot of what we're talking about when we think about MX2 is say more.
What we do is more like let's say back office document operations type work. Right. What I think about often is from a document or any input coming in to like how it gets captured, how it gets routed, how it turns into sort of work and then that last piece of it and the turning it into work a lot. Obviously a lot of that is what MX2 is doing and things like that. So I think it's a, it's an interesting one. What you were going to say something
[00:23:44] Speaker A: though, there's just a. There's a balance between having an opinionated software and a sellable software. And I think those actually are in conflict with each other.
[00:23:52] Speaker C: Think about this, right? Where a lot of different AI platforms in the legal vertical are going to go, whether it's consumer or B2B is actually very customizable because of just the concept of knowledge and skills. So pretty much every firm right now that is on the AI train is either going to try to build knowledge themselves and skills themselves very specific to their organization, or they're going to work with a vendor or an AI company that will help them do it and build an opinion for them or a range of opinions if they don't have the capability of documenting themselves.
So technically that's opinionated, but also that's customizable. So I don't know if it's really going to be this like, hey, you're gonna have a point of view. It's just your point of view is going to be reflected differently. Instead of having very specific set of dashboards, now you're gonna have very specific set of agents that run on a very specific set of cadence that's important to your firm, that is more reflective of your strategy. So I don't think it's one or the other. Does that make sense?
[00:24:54] Speaker A: Yeah. What do you think? Yeah, I have an opinion. I guess I'll share mine. I try to see out of it a little more. But at lexamica, because we're a multi sided platform, we have to be relatively opinionated because we have to take VIVEC and opinionated. Opinionated is there's a way that you do referrals. You can't just come in and tell us like, hey, we do referrals this way. Now we have best practice, best practices. And then that gets baked into our core workflows so that there's interoperability between a thousand law firms. Now if I'm building an AI chatbot and deploying that independently, I don't have to be as opinionated. But on the other hand, like, what I think is interesting is that Vivek, for you, there are probably ways to extract information around documents that are objectively better than the way that Joe, who's been doing this for 20 years, thinks that it should be done. Yath, on your side, you've built the most efficient practice of law and so your software. Two reasons. One is you do actually know how to do it better and people will get worse outcomes if they don't listen to you. And two is a lot of people are probably going to be buying MX2 because they want Morgan's knowledge built into a chatbot. And if they're buying you, they're choosing that over either a generic tool or anthropic itself.
[00:26:13] Speaker B: I get your question better.
[00:26:14] Speaker A: So I think that's what I'm trying to get to is like how opinionated should software be in the agentic AI era and how much should it just map to what law firms themselves are saying? This is my best practice, make it
[00:26:26] Speaker B: So I think that you kind of answered part of the question yourself earlier, like it's easier to sell something that is less opinionated, but then you want to actually get best results. You want it to be more opinionated. And there's probably some degree of opinionation using your term that you need to make it work well, especially with if you're going to be integrating your software with deterministic software that does require specific inputs and things like that to trigger flows in Salesforce. But in our case, we've had to make our software so configurable that it is basically customizable. And you've got to meet customers where they're at. Right. And so I think the hardest thing for a law firm is not buying the technology. It's the change management. And from our perspective, we have to build software that, like, if you want to do it this way, like, within reason, it has to be able to do it. But our job and like the people we have, we don't take requirements and set it up. We are trying to be as consultative as we can be and how you should do it based on your goals and what you're trying to accomplish with your firm, like in your ultimate. Like what? Your ultimate workflow. We usually try to map the entire thing out. All right, this is what you're doing now. There's five handoffs in a document before it gets there. You have different naming conventions. And this and this.
And this is what it could look like to get your end results. You want. Now, not every firm accepts that, and not every firm accepts it, like, at the outset. And so sometimes you have to start them closer to where they are, and then they move them along. And some firms never move along. And some firms are using our software to the fullest, and some firms are using it not to its full capacity. So I think to sell software, though, you have to give them the option.
[00:27:57] Speaker C: This is the difference between building to serve potential clients and then building to serve yourself.
[00:28:02] Speaker B: Yeah, but we talk about places for your attorneys where they can have flexibility on.
[00:28:06] Speaker C: Yeah, for sure, we want to let our best litigators do what they do without interfering in their workflow. But there are certain areas that we just want to be done a very specific way in terms of client experience and managing risk. And for those, we've made investments in a compliance team to make sure.
[00:28:21] Speaker A: Sure.
[00:28:21] Speaker C: It's done a very specific way in which we don't really rely on just technology. We also want people. And hopefully in the future, it'll be more technology that does this. But we want to check to see if people are getting calls back. We want to check to see, hey, did we explore every angle on this case?
So there's certain things that I think we just have the benefit of being one business with one core set of values that we don't necessarily need to provide so much flexibility there. This is just our way. This is just our expectation.
[00:28:51] Speaker B: Yeah. And I think that cascades down because there are firms where they have strong leadership and in a firm like that, they can actually do change management. Well, the main thing required, I think, to buying software is easy now, especially because everyone's bombarded by it. But implementing software is difficult.
[00:29:06] Speaker C: And so implementing is so difficult. And the change management piece is hard because.
And I'm sure you guys see this. You go to a law firm, you'll see that there's a law for boner. There's maybe a COO, but really there's a paralegal who's been there for 20 years and she really runs like things on the ground and she knows all the best practices. And ultimately all these other people are making decisions. But in order to get it done right, you're gonna have to work with her and her team and they should
[00:29:32] Speaker B: bring that person in when they're making the decisions. And also what we see, like, when we see, I won't say failures, but the best implementations are when the executive sponsor stays involved or there's someone who has oversight over it. But what you see often in legal is someone a managing. An owner is pumped. They buy something and they go, great. And they go, hey. And then they bring in the actual OPS person. Maybe they don't have a coo. Maybe it is a paralegal in there, Hey, I bought this. Go make it work. And that person wasn't involved in decision making. And also they haven't communicated with that person the vision they have for the end. So they have to. They don't communicate well. This is why we bought it. This is going to change. Just little examples in our world, right?
They could have a naming convention that they've had forever. It could be a terrible naming convention. And then you're like, okay, the firm owner has agreed now. Or they have no naming convention or different ones. And it's like, hey, you've decided you want standardization. So this is. We're going to work with you to have one naming convention or whatever. And they agree. And then they're like.
And they. And they push back crazily on it because. And for the. I've seen the craziest things. Let's say we. Our software might pull Dr. John Smith, MD. And they're used to putting Dr. Smith. So they're like going in every time and erasing the John Smith and making it what it was. And then you have to get the firm owner back in. And this is a real conversation where they kept on changing this. And then the model thinks it might be wrong. Because it's pulled out the right thing, but then they're changing it so we have to go back. And then firm owner is like, I just had you putting in Dr. Smith because it's shorter and I want you typing less. In fact, I don't want you doing any of this, so let it do what it's doing. I prefer to have the full name. But these little things that they don't communicate effectively across because the good employees are going to keep doing what they've been trained to do or what they think they've done. Well, like the people in the call center that keep on riding on the pads. This is how I do a better job.
[00:31:12] Speaker C: So yeah.
By the way, has anyone in the world thought more about naming conventions than you? How much of your time have you dedicated to like, naming conventions, Dylan?
[00:31:22] Speaker A: Man, I've never had someone say the word naming convention to me.
[00:31:26] Speaker C: I hear it already come up like six times.
[00:31:27] Speaker B: I talk about naming convention.
[00:31:28] Speaker C: That's the entire podcast. Naming convention, naming convention. We're here to talk about AI.
[00:31:35] Speaker A: This guy's just talking about file names.
[00:31:36] Speaker C: Why do you not think I run a real AI company? Because you spent all your time thinking about naming conventions.
[00:31:41] Speaker A: I'll give you a chance to clap back because, Vivek, I want, I want to push back, have you say, look, they've spent a bunch of time, they're building this MX2 thing, but they are still building for one customer. That customer has many users, but what starts to break when they roll it out to other when they get outside of Morgan and Morgan when they can't walk down the hallway to an engineer and go fix something live inside of Morgan, what starts to break faster? Like, it's like you have all these best practices and this beautiful product and the Morgan team adopts it and you roll it out and then your first customer says, I don't like the colors. Right. Or whatever it is. Or now you're dealing with third party vendors that you have to go integrate into and systems that don't want to talk to each other and all this crazy stuff. And you guys have been down the road, you built, launched Solidify and that wildly successful. So it's not that you don't know how to do it, but I think going out from Morgan, where it's in some ways it's a very friendly environment, people are yelling at you because they want to do their jobs better. The next people are yelling at you because they paid you money and now they're pissed.
[00:32:40] Speaker B: I think building software is hard. Deploying software at scale is much, much harder. Now Morgan, obviously it's a different animal to ask about Morgan versus asking. I mean Morgan's committing a billion dollars into, into building this. Right. They're spending a tremendous amount of money and I know their team, they're building it like a best of breed startup
[00:32:56] Speaker A: would build this technology, but they're still building it internally. And I'll push on this because there's been a lot of big investments that have failed because step one works and step two is a separate set of problems.
[00:33:08] Speaker B: Yeah. So my point isn't that I know that MXQ is going to succeed. My point is I know that they're thinking about it, at least from the perspective of building it as now, at least revamping, maybe it wasn't always, but looking at actually making it productized in the right way before an enterprise grade, before rolling it out. I think what's more, what I see more of, and I'm much more disparaging of, are smaller firms that vibe code something for their own firm.
[00:33:33] Speaker A: We're going to get to that. We're going to get to that. Vivek, stay on topic.
[00:33:37] Speaker B: Well, you're asking me to on Morgan's
[00:33:38] Speaker A: neighbor, I'm saying step one is really strong. How do you prep for step two and what are the failure modes and how do you get ready for that?
[00:33:44] Speaker B: Yeah, yeah.
[00:33:45] Speaker C: In terms of failure mode, I think you're pointing at an obvious one, but I think it's solvable, especially that look, technology with AI is easier to build than ever.
[00:33:54] Speaker A: Yeah.
[00:33:55] Speaker C: So building is not the problem. It's managing people, it's managing relationships, it's making sure you meet expectations. And that's more process oriented, that's more about how you manage clients, what their expectations are, what they want their specific roadmap to be and communicating that effectively. I think, and this is a cop out, but it's also the truth because that is where really the magic is. If you can manage those relationships and you can manage those expectations, you're going to be able to deliver. And whether that's internally at Morgan or externally. And I've worked at a software business before, like I know how this works.
It's the same thing. It's the same thing. But the key thing most people want is someone who's going to build something great, someone who's going to commit to delivering it on time and then who's going to support them in terms of the change. And there's certain things that ultimately it's going to be on the Law firm operator and the team internally. You deploy a product, of course you can measure it, but in terms of helping them on the ground, talking to people about the why, people often need to know the why. If you go to a paralegal and say, we have AI now and by the way, we're going to automate all of you, you think they're going to be excited to use it? Like they're not excited to use it, but are you.
[00:35:06] Speaker B: And I guess the question I have before, are you going to split out, you think a completely separate division to sell, support, implement MX2, or will it be like Litify was or.
[00:35:17] Speaker C: Yeah, I think there's, I think a few things. One, we're really excited about Amex 2, but Morgan, and Morgan is the heart of our focus. Yeah. So we want to make sure we're able to support the needs of growing Morgan. And that's like number one priority. Yeah, number two, I think in terms of focus, it's not going to be sustainable for us to have the same team do two things and you have
[00:35:38] Speaker B: to build specialization and that's the main thing. So if they do that, it's, it's essentially. Actually that's why I think of it more like they're incubating an actual tech company within a large installed base and then separating out. Now, if they tried to run and sell MX2 like out of Morgan, then I would tell you a million failure. You can, we can all see the failure that would happen there.
[00:35:58] Speaker C: And I think this is where MX2 goes. And I think it's going to be important for law firms to own their own legal intelligence. They don't want to be on Morgan's tenant, they want their own tenant, they want to manage their own data, they want their own experience.
And I think there's opportunities to give it to them. I don't see it like, hey, you're now going to use Morgan's platform, but I think we have a long way to go. I'm most excited about what we're doing at Morgan.
[00:36:21] Speaker B: Yeah.
[00:36:21] Speaker C: Because I'm seeing the returns, I'm seeing the opportunities. Yeah. And I think there's even more for us to do that we're really excited about.
[00:36:31] Speaker A: Let's look externally from this conversation for a second. I'm really interested to hear your take on this. Yath and Vivek, I'm sure you have a perspective too. Where are the insurance carriers? Where's the, the defense on this? My, my take is that the PI side is just crushing right now and I think that five years ago, there was this kind of a sense that the insurance companies had more money, more technology, this idea of Colossus. But as I see it now, I think that you guys are. I think we're just running ahead.
[00:36:58] Speaker B: I started an insurance defense, and we actually started our business like selling to insurance companies. Anyone who was pushing that narrative is an idiot. I never thought that that insurance.
[00:37:08] Speaker C: I didn't start in insurance defense. When I got to Morgan, I heard rumors of all the tools and technology that the insurance defense firms like Colossus, was a scary thing. Now I've learned that maybe it's more of a scarecrow than a scary thing, but I don't know where they're at. But I do keep in touch with people who've been in the legal tech community who built products in both spaces. And I was talking to someone who's really smart, who's at an insurance carrier, and he basically told me that they're very behind. Yeah. And I think the reason that they're behind is actually structural. Yeah. The thing about most consumer law firms is you could actually make decisions pretty quick.
[00:37:47] Speaker B: 100.
[00:37:48] Speaker C: Hey, I want to work with this vendor. I want to do this security review. I want to. You can move at a pace that will allow you to integrate and adopt technology quickly.
[00:37:56] Speaker B: Even a Morgan will move twice as fast. Way one, twice fast as a smaller. As an insurance company with less people than they have. And that's exactly right. So selling into insurance, like years ago, our first customers were insurance companies in the early days, in 2018, 19. And the innovation department would be like, we love this. And come in. Then you find out the innovation department has absolutely no pull within the organization. IT department's like, I don't have any bandwidth to commit to this for two years and my roadmap is not. And then internally, they had so many systems because they bought. Here's the thing. Because they bought technology first, they got leapfrogged completely. Just so you don't see wired telephone systems in Africa and parts of it. So they all bought a bunch of legacy software. They built a bunch of patchwork IT integrations across it to make it talk. Built these massive departments with IT people with these fiefdoms to keep it all together so that when new things came out, it's okay. Like, it doesn't integrate. You have to build custom integrations to everything.
So they're so slow in their bureaucracy to move that they just stayed way behind even if they bought technology.
[00:38:53] Speaker A: And I think that's being exacerbated because the rate of change in technology has increased over the past five years. Like you said earlier, like you've been able to rebuild this thing three times in certain areas in a couple years and insurance companies haven't even bought the first version of it yet.
[00:39:07] Speaker B: It's important to see though, there are now like insured tech startups and things like that are taking this AI native approach and coming in and working as teams. Yeah.
[00:39:15] Speaker A: But that doesn't mean that they don't have full adoption yet. I just, I think, and this is exciting to me because this. We're on the.
[00:39:21] Speaker B: We're.
[00:39:22] Speaker A: I believe we're on the side of the good guys like that. Like there is now the balance of power at some point will shift even though you. Because insurance companies are huge, right?
[00:39:31] Speaker B: Yes.
[00:39:31] Speaker A: Then they have a lot of regulatory capture, they have a lot of lobbyists and I'll say all. And a lot of money. But what they can't do is move quickly.
I think that's a really. I'll bet on someone who can move fast over a legacy incumbent.
[00:39:44] Speaker B: Yeah. One thing also I'll say is which surprised me the most because like I started my career in insurance defense and insurance defense, like firms, like they're all kind of fighting for the insurance company scraps. So they're all competition. They don't share information, they don't share best practice. You'll never have a RevCon conference where the best insurance venture goes and teaches people how it does things. Or all the conferences that exist in the plaintiff space, people are sharing information. Or you go, like in California where I live, and you have Cal and you have these organizations, like, they don't have that. And so the plaintiff bar has always been much better because on the side of fighting for the right guys, like sharing information in a way that has just made them so much better at adopting technology. So even when one firm gets technology, like even what Morgan's doing, it gets disseminated across technology knowledge. Experience gets disseminated across.
[00:40:35] Speaker A: It's still a rising tide. It's still a rising tide when Morgan is sharing what they're building. Because whether someone else is building something, buying something off the shelf, building internally, ultimately getting MX too, it doesn't matter because the rate of change is faster on this side.
[00:40:49] Speaker C: I'm always interested in systems and for me it's actually interesting just how like wanting to build technology, knowing what technology to build, and ultimately having the capability to build technology doesn't matter if you don't have the structural requirements to be successful in that because, hey, can you get approvals? Can you do security reviews quickly enough?
Can you get pilots approved?
Those are all like things that just are super important. I was actually talking to gc, a really large company, and she's smart. Her workflows are ripe for automation and it would add so much value to their business in terms of just capability. She has a team of 20 people. They're all overwhelmed. They use a lot of outside counsel, but she can only use two vendors. And one of those two vendors is, I think, SAP. You want SAP to implement your AI? It's not the right structural environment to succeed. And I think for these companies, they have to have change on that front. And then also I think the incentives are there. The. The defense attorney.
[00:41:59] Speaker B: As long as they hold onto the billable ally, there's no incentive.
[00:42:02] Speaker C: Yeah. They're not necessarily going to be so incentive incentivized to want adopt AI and then two actually be super transparent with the insurance companies either. Yeah. I think that really.
[00:42:13] Speaker A: Yeah. There's a slightly. And a slightly counterparty relationship between defense and insurance companies that doesn't exist on the plaintiff side.
[00:42:21] Speaker B: Yeah, they're. Because they're not. They're interested in misaligned incentives. Completely.
[00:42:24] Speaker C: Sure. And I think that is a huge opportunity for actually technology to resolve. But I don't think anyone's been able to make progress there because the. I think the insurance companies can't get anything improved and the defense attorneys probably don't want more accountability.
[00:42:38] Speaker A: Yeah, it's difficult. And I think that I'm a big believer in aligning incentives. At the end of the day, when you have the plaintiff's attorney and the client aligned around maximizing the outcome in the shortest amount of time possible to get the most amount of recovery, that is a really strong incentive alignment to build technology that both saves time and also increases the outcomes. And then that's why. It's another reason why I think that tech's pulling ahead. We gotta land the plane here.
[00:43:07] Speaker C: I think tech is a huge thing, but I think the regulatory capture piece is also important because that's a different game that can be played by them where maybe the plaintiff's side doesn't have the same structural advantages. Right. The PI law firms more have a structural systematic advantage where they're able to make decisions quicker, move faster, adopt technology at a higher rate, become more efficient. But in terms of regulatory capture, they might be able to do more in terms of alignment with political parties in terms of raising more money, corporate sponsorships. So it might not be a game where you just win by being efficient. Right. So it's multidimensional.
[00:43:47] Speaker A: That's absolutely true.
You can lose the game by looking at the wrong part of the board.
[00:43:53] Speaker C: Yeah. They can just change the rules. They can just change the rules. And in our experience, they've tried to change the rules. We have a focus on, you know, publicizing that and making that transparent to everybody. It's making sure that the rules are actually good for consumers.
[00:44:07] Speaker A: That'll be the next podcast. What is something that will be true by the end of next year that firm owners need to be thinking about now? I'll hand it back to you, Vivek, and then you yath and we'll end there.
[00:44:18] Speaker B: I think that consolidation is something people should be aware of, whether or not they want to be part of it. It's going to ensure that to compete, you have to be more efficient. Just to compete and continue competing as a firm and stay where you are. You're going to have to get more efficient because the whole industry is going to start thinking about operations a lot differently. It already is.
[00:44:38] Speaker C: I think you'll actually see long horizon, agents and agents doing work. I think everyone is talking about it, at least outside of legal and in legal. I think the legal tech companies are talking about it. I think the implementation is nowhere. I think the implementation is nowhere because of a lot of the things that Vivek's highlighted on the call. People really don't know what their workflows are. There's bad data. Yeah. And then there's a ton of institutional knowledge that's not documented. So I think there's this case companion boom that everyone initially saw with AI that's like the outside of foundation AI, which has been around for years. Apparently it is the first AI company in the world.
Vivek invented the concept of AI.
[00:45:23] Speaker B: I invented AI. I am AI.
[00:45:25] Speaker C: No, to land the plan, I think we had this first wave of post chatgpt AI that I think is successfully being adopted in consumer law. And I think the second wave is agents. Everyone's been talking about it, but I think in that time horizon, I think you're going to succeed. Successful implementations and people actually getting value from that. And I'm pretty pumped about that.
[00:45:47] Speaker A: Awesome. Yeah. Vivek, thank you both for joining me on the podcast. I think we crushed it today. Really good stuff.
[00:45:53] Speaker B: Thank you for the new format.
[00:45:55] Speaker C: Yeah, yeah, yeah. Great to be here.
[00:45:57] Speaker B: Yeah.
[00:45:57] Speaker C: So much.
[00:45:58] Speaker A: Thanks, guys. Join us next time.
[00:45:59] Speaker B: Thank you.
