Why is Harvey a Michelin-star restaurant, Claude a McDonald's, and Ruli a neighbourhood boutique — and which audience does each one actually serve?
Bryan Lee, CEO and co-founder of Ruli AI, breaks it down for us. In 2024 he co-founded Ruli with Xi Sun, formerly a machine learning engineer, to build an AI platform for in-house legal teams that focuses on continuous legal intelligence rather than chat wrappers or case law research.
00:00The financial models are more focused, driving enterprise license volume per year, 50,000 person team, rather than the 100 legal licenses that are part of that company. The way that Harvey sort of curates things sort of like a Michelin star experience. A lot of hype around it and for a certain type of audience really want to get behind that curtain and see what that's all about. There's minimum seats, minimum spend, all that sort of stuff. And then you have us. I call it kind of your boutique neighborhood restaurant. It's a place where the owner and the chef will come
00:30out and remember your name, remember your kids' names. They're going to take requests from their customers and really deliver that really great neighborhood family experience. Most of our customers are looking for that magic button.
Speaker change: But we're getting dangerously close to letting the users press that magic button without knowing how to code, without having any specific professional knowledge. Don't you think that is a big threat to all tech companies out there?
Speaker change: Yeah, it's a great question.
Speaker change: And Epic just moved into legal. So what happens to all the other startups
01:01building in this space? Today, I'm speaking to Brian Lee, CEO and co-founder of Ruly, a legal intelligence platform for in-house legal teams about exactly that. Brian and his co-founders come from a big tech and big law background. And in our conversation, I ask him how he translates that experience into what he's building and what his read on the legal tech market is now and in the future. Let's get started. Brian, so happy to have you with us today from Texas.
Speaker change: Yeah, Anna, thanks for having me. I moved here actually, you know, a few years ago from the Bay Area.
01:31Speaker change: So why did you choose to move to Texas and build Ruly out of Texas? Cuz my understanding is that a lot of these legal tech companies from the US are either based in San Francisco or New York, but you you chose Texas.
Speaker change: Yeah, I mean, I think for me it was you know, a bit of a personal decision to move to Austin first and then the legal tech came after both my co-founder and I left Meta Reality Labs building AR glasses. And uh co-founder is actually based in San Francisco. I think you get a better perspective uh when you have a
02:01remote-first team. And we have some folks in like Colorado, for example. And because the Bay Area uh it tends to kind of uh be you know, subject to groupthink, which can be like, you know, an accelerator, but can also be a limiter in a lot of ways. And so, I think it's great to kind of get that different perspective. You know, you go to different parts of like, I mean, Austin or even, you know, as we you know, begin to other other areas like Nashville, Atlanta. There there are people who've never heard of Harvey.
Speaker change: Tell me a little bit more about the founding team of Rolly and how you work
02:31together and what kind of dynamics you have as a founding team?
Speaker change: So, my background is in uh you know, computer engineering and law. Did big law in Harris. And then, you know, found out that, you know, I'm really more of a product product guy, ultimately. And built a lot of products from the first Google Assistant, working on that to like early LLM transformer technology at Google and then later. And then, my co-founder's experience, he's he's been, you know, he held a number of machine learning patents that he's been building
03:01the uh machine learning and pre-cursor AI, you know, at Amazon, you know, a decade ago. And then, has that experience through um places like LinkedIn, Airbnb, and also Facebook. You know, we have a GC on team.
Speaker change: So, Rolly today serves a lot of smaller legal in-house teams. And from our conversation thus far, you mentioned that the core founding team all have big law, big tech backgrounds. So, what makes you think that you would understand the problems and the challenges of the smaller legal teams
03:32given the core founding team all came from these big companies and big law firms.
Speaker change: Yeah, I That's a great question. I You know, I think like what we found is uh you know, our focus is on mid-market at the moment. I mean, like we also have like larger enterprises in our pipeline, which is great. And I think they they see some early successes and have have come knocking at our door, you know, but what I think what's also surprised me is actually the you know, legal teams are actually quite lean. And so we have a you know, a sole GC
04:02that's also part of our you know, customer base that works for it, but he's sole GC of a public company that's listed on NASDAQ. And that's something new that I think that I maybe not had not realized before coming into to Legal AI that actually you know, companies can be quite substantial, but legal teams are still quite lean, right? Not Not every legal team's going to be at that sort of base tech level, but actually a lot of them are in this sort of two to 25 mix. And you know, we call it a market, but actually a lot of the companies are are you know, I guess punch in a higher
04:33weight class than you'd expect.
Speaker change: And so to like maybe help educate me and also other folks who are trying to understand, you know, what are the different needs of these teams if you cannot segment by the size of the company because even like you said, NASDAQ listed companies can have very small and lean legal teams, whereas you know, a mid-size company could have potentially a bigger legal team. Like how would you bucket the different needs of these legal teams if it's not by, you know, how big the company is?
05:03Speaker change: You know, ultimately I think it's a a similar problem problem sets that I think legal teams of all sizes share that we've grounded on. One is like how do we you know, move at the speed that the business moves, right? I think that's pretty universal. I mean, that could be more efficient. It could be looking at things that are more self-service for the business. Two is like there's this need to have maybe reduce outside counsel spend. And I think you know, if you're a smaller team, it's it's almost like yeah, there is no
05:33outside counsel budget actually. And I think when you look at the the very large teams, you know, sometimes it's like wow, our outside counsel budget is you know, tens and 50 million dollars or whatnot.
Speaker change: Especially you work big companies.
Speaker change: Exactly. And and and the legally eyed products might only be like the cost of one head count. And so it's you know, the the economics are are are are really sharp. And I think that, you know, I think there's also this this third area that that we hear a lot, which is um you know, everyone's sort of dissatisfied uh
06:04with their CLM.
Speaker change: It is very confusing for a lot of legal teams as they're also trying to navigate what kind of problems they tackle. And I'm sure it is for the founders as well cuz there's so many different directions you can go in terms of building your product. And from my understanding, for a legal intelligence platform category product, obviously everybody from day zero started out with a chatbot and then expanded beyond that. And then we had the integrations into Word um and all those gradually all became table stakes. And then you make a different direction call on which point you go to. One is I
06:36go to build a CLM. Another direction is I become the legal front door. I help build out the business self-service route. So how did you grapple with these different directions and which direction um to focus on?
Speaker change: Yeah, no, it's a great question. And I think one thing that we actually want to call out is uh we didn't start with a chatbot. So um what's interesting is I think I think the majority of folks did and I think for us, you know, summer 24 where where we initially uh started, uh we actually
07:06started with an agentic vision. But the market wasn't really ready for it. Uh so originally we were building a kind of like a reinvented, you call it legal front door, but with a digital paralegal that you could really orchestrate and and and work directly with you, but also for the the business and sort of do that triaging self-service to also automated review. But I think at the time what we what we found and and also it was you know, grounding your institutional knowledge. And what was interesting at the time was uh legal teams are just asking for a legal co-pilot. And so uh
07:37they're like, we love the idea of you uh grounding our institutional of We don't see that with other players out there. So, how can you add all our contacts of our company and give me this sort of custom legal AI assistant, but I just I just want to talk to me, right? Cuz there's this comfort level of like and then you didn't you know, the word agent wasn't even out there yet, right? So, we were calling it digital paralegal cuz the word agents wasn't even in your vocabulary yet. So, then we actually built more towards this co-pilot vision, but always with a platform in mind because we thought this sort of you can
08:08call it cloud legal moment was inevitable. And I think like you have to you know, I wrote posts about this recently where you have to understand are you building in the jet stream or not? And what do I mean by that analogy is like the core foundational models are sort of like, you know, this you know, these fighter jets are like launching this jet stream and you don't want to be caught in that jet stream or you're going to be eviscerated. And I think like, you know, one of those examples is like, you know, some of the coding tools since they've built in the only act direction to really take on coding as one of the higher use
08:38cases. If you're simply coding agent, you're going to be in that jet stream. It's okay to be in there, but you have to recognize that you are. And then for us, what we realized was if you're just building sort of legal AI to more of a chat wrapper, eventually you're going to be eviscerated in that jet stream. And so, we've always been having this platform vision. And I think what's exciting for us is, you know, we we build an area where we have legal assistance focused on research. We do contract intelligence but a way that focuses on post-issuance intelligence. We're really deliberate of not necessarily swimming in that same lane
09:09lane even though we've had some asks about it and not doing that free searching to workflow work. You know, maybe someday that could be true, but for now we've we've delineated to be more of this you know, continuous legal intelligence and giving you sort of agentic insights all the way. And then we of course we're in the word plugging as well and we're stitching that vision together. That allows us to serve not only commercial counsels but product counsel. I think a lot of customers what they love about us is that we're able to serve all members of the legal team in a way where they're
09:40able to buy one platform, grow with it, and have different members of the team really benefit from some some part of Ruly.
Speaker change: Yeah, so I I think that's super interesting that instead of like building deeper into the like niche of contract automation, which then you would, you know, try to build out a CLM, a workflow management tool for, you know, the contract life cycle process, you decided to move horizontally and try to also build out solutions for different types in-house counsel within like company. So, like product counsel
10:11is one of them. It would be really interesting for me to hear an example of how Ruly serves and helps, let's say, product counsel or employment counsel in a way that maybe some of the other platforms don't do at the moment.
Speaker change: Yeah, it's a great question. I think like for for us, what we're able to do is, because you're able to build it, bring your cloud drives in, and search that search that information that you have in past. We have people that will use it to help draft new policies, for example. Like, there's product counsel who needs to draft new policies, as well
10:42as things like looking at, you know, terms of service, but while while incorporating that with their legal research needs. The same thing we're seeing with employment as well, or more regulatory lawyers, looking at, you know, the regulatory or or employment requirements across. One of our, I think like one of our customers is Sequoia, that does HR benefits, and they do exactly that. That's the nature of their business, and they use a lot of our our product to stay ahead of regulations, compliance in the employment space. So, they're able to actually bring that into their
11:12core core product offering itself, but then as well as how do they benefit from a platform to really also serve commercial needs as well. And I think where we're going with my earlier point was, I think what's exciting for today is I think we're moving back to that angentic sort of vision, potentially, because I think right, you know, what we're seeing now from a lot of customers is I think they're very comfortable with this co-pilot platform experience and we're in that phase next where
11:44you know, there's there's there's a lot of lawyers asking for how do we automate workflows and build agents and you're seeing that curiosity now and I think we're at that phase where we can start realizing, you know, what has been our long-term vision of sort of moving, you know, in this orchestra analogy uh the violinist to be sort of the conductor, which is also like, you know, kind of the long-term vision really is having, you know, different agents being those for people playing the instruments and eventually moving that the lawyer in the individual seat to that conductor seat.
Speaker change: So, you mentioned earlier about meeting
12:16the users where they are, which is back in 2024. The users might not be ready at that time for agentic workflows, but do you think it could also be the case that even if the founders had the vision for agentic capabilities and workflows, the tech just wasn't there yet because I know at that time, you know, even like with chat queries, there were a lot of hallucinations. I remember attending a conference even in 2025 where conference where someone in a panel would make a joke that ChatGPT would get the time
12:46wrong when he asked the time. So, like you don't see those kind of mistakes as frequently anymore or like, you know, they can't spell strawberry how like correctly. So, do you think that is the combination of you know, the users not feeling comfortable with that idea and also in addition to that like the model capabilities just wasn't there to realize that vision at the time?
Speaker change: Yeah, no, yeah, that's a great question. I think it is it's definitely part of it. I think the more recent deep reasoning models that we've released that we've also incorporated in our product in the last few months takes us
13:17to a higher level of confidence that we didn't see in sort of that that 24 period. I think so, definitely think it's a combination of sort of like user user trust and option where the technology was. But what's interesting is, you know, we do have early customers using this this module of ours, which is this a genetic product that I discussed that we have it sort of like a secret menu. I guess if you're like if you're at In-N-Out in in the US, you can order sort of order a secret menu. Not hidden if you know what kind of like the the
13:48right way to order it. And we we we actually offer that to our customers and and kind of the irony is it's it's in our first product you can think of it as sort of a legacy product, but ironically it's now coming back as in vogue sort of as like the add-on, the special secret menu add-on where people want basically FAQ answering in Slack or some sort of automation automated review, which we do for some of our early customers directly for the business for compliance review. And I think that's something that we're really thinking hard about on how we
14:19might redevelop that in the second half of the year with some more of the deep reasoning models that that have that have been recently released.
Speaker change: Tell me more about the secret menu of automated review. Can you give me a concrete example how some of the legal teams are using that function?
Speaker change: Yeah, I think it's it's largely in two parts. So for for a public company that we've been working with for a year now and they've decided to renew with us this month which is fantastic. We basically take one of their
14:50compliance workflows where there's documentation submitted in from the business. They're in the real estate business and we we take a sort of like a rolled derivative of like the Fair Housing Act policy with kind of how they interpret it of where they think, you know, high risk issues are, where they would auto prove it, where things would be escalated onwards for for some human level review. And then our effectively we auto prove it directly for the business. In other cases it's pending
15:20and then the you know, legal goes in to to to sort of make any adjustments that they like and review it. But that's you know that's happening you know we probably had processed over like 20,000 requests at this point over the course of the year. And I think that you know things like that are very interesting. I think you know you can apply certain things to any NDAs or marketing compliance review which are interesting use cases that are high volume you know low risk especially in terms of like you know for NDAs in terms of like actual you know in fact for
15:50enforcement and I think a lot of those those things are interesting to explore. And I think you know other ways as well as is really the FAQ answering you know in Slack policy answering and having a trusted agent doing that sort of work that it can program it give it guardrails have it point to certain policies and and that's something that we do as part of our secret secret menu product and we're we're thinking deeply about the evolution of that because we think the market is you know finally getting there in terms of like understanding how you kind of program
16:21agents and orchestrate and I I think that's a really exciting future. At the moment I think like we're we're kind of learning we've learned a lot and I think we want to kind of bring it to prime time later in the year and kind of bring out you know kind of lift the curtain off of the secret menu. But it's been exciting to kind of explore that with some of our early customers who who really you know I think we're early on what agents were and orchestration was before those became coined terms in the industry.
Speaker change: Yeah. So speaking about legal research
16:51cuz we you mentioned it quite a bit earlier and I I've been trying it. It's something that not a lot legal intelligence platforms currently offer or do it very well and I know that really has indexed the primary sources for regulations within the US but I don't see any case law. So I wonder you know what is the product decision around that and maybe talk me through the challenges building out a really good legal research function. Because if it was easy, I'm sure a lot of the other
17:21platforms would have already done it.
Speaker change: Yeah, that's a great question, yeah. I think it was possibly this is grounded in my own sort of lawyering experience. I was never loved Westlaw or Lexis, and I was never litigator. Originally, I was securities lawyer and then later a commercial lawyer in in-house. So, case law has never been my forte. And I think for us, what we realized early on was when we interviewed a lot of in-house attorneys and in my own experience, case law was not an area of focus for in-house teams. You know, they're at a
17:52policy focus on the regulation, what the laws are. They reference law firm memos for certain interpretation of call it landmark cases. But for litigation particular, and we don't steer in that strawberry space for litigators, that might be a subset of very large legal teams. I think that's a a very different workflow. And for the majority of our mid-market, it's it's rare that you have a good litigator on your team. And if they do, that that's not an area that that we serve. And so, we made that deliberate choice. And it it allowed us to focus,
18:24right? And then we can also build our tech stack around the institutional knowledge, which is already a ton of information to get your hands around at the company and optimize on that to give you basically the ability to search your knowledge or just have it as context for the assistant. So, it'll know all your board document information if that's what you've kind of pointed at. And then you're drafting new board materials that can use the basis of that information you passed, for example. And then one of the challenges with case law that I think for the for the non-lawyers that
18:54may be listening is like you really have to look at, you know, sort of the the Shepard's index or whatnot. It's not just about what the case law says, right? You have to look at hierarchy, the rankings, etc. And you know, when I was a trainee, I looked more at the editorial content, which you know, Lexis and Westlaw has written decades on. And I think that's why people are surprised when they realize Harvey actually licenses the Lexis Nexis database. So, they've all this capital, um but they themselves don't, you know, do that case law depth on their own. And I think, you
19:24know, like it that if you're building in that space, it's certainly an area where you not only have to build going forward, but you have to figure out, "How do I, you know, get all the precedent right on all the past, right?" Which is very, very different. And then I think when you look at the technical side of that, you know, if we even if we did all that, um because there, you know, there might be a minority of customers that like that, um that's too much information for AI. So, I think about that, you know, that RAD structure of of how people look at, you know, the, you
19:54know, how information retrieval works at all labs. And my analogy is AI is far more accurate when you're shooting fish in a barrel. And that's why the way that we do it, we curate, um now primarily legal sources for US, but we also recently had a UK EU. And then we also do a curated web layer, um where you're not going to get things from like Reddit and things like that. So, it's really different from like a Google search. We have certain priorities on certain dot gov, certain EDU sites, like things from like Harvard, Cornell Law, etc. But also the Am Law 250 law firm memos sort of
20:25like get published. And with that, you get a far more curated experience. And I my analogy for it is this actually shooting fish in a barrel. And that gives you more accuracy, less hallucinations, more of the things you're looking for, rather than if I just tried to get everything in there, including case law, etc. And then there'd just be too much noise.
Speaker change: So, I really like that analogy, shooting fish in a barrel. I'm definitely going to use it somewhere else. Before we move on from legal research, you mentioned this curated web layer. And I'm sure some people who are Claude and, you
20:56know, Manas enthusiasts are going to say, "Well, you know what? You can do that with Claude as well. I can just, um create a skill to say, you know, only search for primary sources or law firm articles, and then pair that with an N A N workflow that automates and then it sends me updates, you know, the kind of areas of law I'm interested in and send that to my inbox on a weekly basis. What do you say to that? So, I I think like two questions, just so it's fair. You know, for what kind of users would that actually make more sense for them? And then for which group of users it would
21:26make more sense to choose a platform like Rulings for this kind of task?
Speaker change: Um, you know, it's a great question. I think like with, you know, our platform and our workflow and the way you can DIY and build skills, I think there's, you know, it's it's really coming coming down to kind of like a DIY versus, you know, buy buy an enterprise SaaS product type of experience. And I think that's, you know, a place to keep in mind and and to really, you know, think about that cuz I think there are a lot of
21:56things that are beyond sort of the product itself, which, you know, as a as a purchaser or consumer, like you're you're buying into things like, you know, customer success. You're buying into things like the depths of the quality of like the UX features, and how deep can they go, right? Ultimately, the the the foundational models are more focused on, you know, driving just enterprise license volume, right? A lot of volume comes from even 50,000 person team, you know, maybe 20, 25,000 could
22:26be engineers like in a tech company. That's really their focus rather than 100 legal licenses that are part of that company. They're using that more of a differentiator so they can be many times your other foundational are really win that whole package, right? And I think there's a level of focus that they won't have necessarily going in that depth. And I think if you're a customer of say, like let's say you're a legal customer and you're trying to get, you know, your priority built of like what what your your feature requests are, it's highly unlikely that with these foundational
22:56models that you're going to have a customer success person that you can even talk to depending on how large your team is or if you're just, you know, having a subscription model and there's only a couple of you on team, you know, highly unlikely. And then too, like, where is that priority stack? You know, they're probably going to build for the the the coding agent more likely first or, you know, the largest, you know, the largest voice within that ecosystem. So, I think all that gives you a very different level of depth and experience, but it's up to you to decide what that level of depth that's appropriate for your business and
23:26for your needs. Yeah, I have a great analogy for this. I think you know, kind of put things in perspective. I think so we use like a sort of like a food analogy, restaurant analogy. So, I think, you know, the way that Harvie sort of curates things sort of like a Michelin star experience. It's a lot of hype around it and for a certain type of audience it's like, oh, I really want to get behind that curtain and and see what that's all about. And then there's, you know, big velvet rope, you got to like book your book your seat, you know, 3 months in advance, put a deposit, you
23:56know, there's a lot of restrictions, you know, there's minimum seats, minimum spend on that sort of stuff, right? And then I think you have your financial models through your, you know, you know, Cloud, etc. That's sort of like your McDonald's. I think it's like mass mainstream. It's like, hey, you know, you have your menu options. It's very quick serve, but like you're not going to get, you know, any real sort of depth and customer experience or you're not going to be able to, you know, affect change or like get your, you know, your sort of products, you know, your your product requests sort of met necessarily and influence
24:26that road map. It's really about scale and and speed. I think you have, you know, us and maybe some others in this I'll call it kind of your boutique neighborhood restaurant, right? It's a place where, you know, you can reliably know that they're going to they're going to they're going to be there, you know, as part of the community, you know, the owner and the chef will will come out and remember your name, remember your kids' names, right? They're going to shape the menu in a way that could be seasonal and they're going to shape requests from the customers and really kind of deliver
24:58that really great sort of like neighborhood boutique style experience. And I think ultimately, it depends what you're looking for. And you could potentially start with sort of that, you know, mass McDonald's experience where you you you tried that, you weren't you're curious. We see a lot of customers curious, "Hey, I want to see what Hardee's is about." And then some of that flashes and leads maybe some disappointment about expectations, but I think there's um you know, value to to kind of what we are offering for the right audience where there's this boutique experience where a lot of
25:28customers say, "Uh we win um because of our customer success sort of boutique white glove experience that we deliver not only at the uh you know, that we start at that pilot stage um that we bring all throughout the the entire customer journey with us."
Speaker change: Yeah, I I thought your comments uh on LinkedIn for my cloud review post was really amazing. The painting that bigger context of, you know, Anthropic's broader strategy of, let's say, why they would launch, for example, a word plugin. But this analogy with the different restaurants, I think is even
25:58better. It's hard to top that game. But you do wonder though, like some people could say that maybe in the future you don't need to have that custom white glove service because the foundational models um even if it's quick serve, even if you can never, you know, have your own personal customer service agent like a account executive, the models and the skills and the whole plugins will be so powerful that you can just DIY and create whatever kind of experience you need, and that will be enough for most
26:28users. Do you see a future where that could be possible?
Speaker change: Not everybody wants to do DIY cuz for a couple reasons, I think like ultimately, most folks, you know, don't really are not going to dive into that DIY necessarily. Most of our customers are looking for that magic button. They don't even necessarily want to prompt that question. I think like where where it is important though is to think about differentiation in your product because I think what's clear is if you are only sort of doing chat necessarily and it's very surface level chat like maybe there isn't deep legal database resources or
27:00sensor curation that gets very vertical or if you're only doing very surface level work plugin and you're not going into that depth of like surgical redline or time intelligence backed like everything that customers create for in like it's your intelligent posture etc. Then yes, I think that gets eroded. I I do think that gets eroded, right? You know, is are they going to build you know, e-discovery tools and all those other things? You know, it's I think it's it's a harder stretch to think that everyone's going to DIY everything based
27:30on based on a foundational tool.
Speaker change: So you say that not everyone will want to DIY, they will just want to press a magic button, but we're getting dangerously close to cloud code and cloud code work letting the users press that magic button without knowing how to code, without having any specific professional knowledge base to create those DIY solutions that seemingly meet their needs. So don't you think that is a potential big threat to a lot of legal tech companies out there?
28:00Speaker change: You know, I I think you know, something I think about is you know, I think there's a distinction of moving legal folks from sort of the hot seat to more orchestration, but the distinction being you know, how many of those folks want to move into that product builder space. And I think when you look at that broad market, I think it's you know, I think the early adopters and being at early adopter side, you have more folks wanting to do a DIY and that experimentation. When you start talking to folks
28:31in the broader landscape with maybe maybe Texas and and having a different purview, I guess the majority of of the way that we're seeing sort of that purchase cycle and and that intent is more folks are like, "Hey, we want to trust you that with Ruili and their expertise on the tax side and legal side that you've kind of curated the best experience for me. So, I think there's this idea of curation that's always been in sort of the human evolution that
29:01people really like. And so, I I think ultimately there'll be people that are will be looking for some sort of curation, for some sort of expert, whether that is from us or maybe it is from a DIY curator expert on their team. But ultimately, I think there's somebody there where that curation is part of their DNA and other folks are like, you know, like I I think I want to trust someone else to do that curation for this particular domain versus you know, other you know, other specialties.
Speaker change: Yeah. So, moving on from the foundation
29:33models, you mentioned this curation process as being very important for a lot of legal teams. So, what would be your advice for legal teams that are exploring different legal tech options and each option like each vendor tells them that they can offer a curated experience for them. What do you think are some of the good proxy metrics that legal teams should be looking out for when they're evaluating whether a vendor is a good fit for them?
Speaker change: Yeah, it's a great question. So, for us I think often, you know, what we've seen
30:03is legal teams just kind of making a list of, you know, maybe like legal AI vendors and plus your LMs in that mix and just creating a mix of all sorts of different things that do diff- different things in different categories. I think you know, you want to start foundationally on like looking inside and figuring out like what is it that we want to solve? You know, maybe going back to those earlier buckets. Is it some sort of efficiency for the business? You know, what areas, you know, as a business do we focus on? Do we do more contracts? Do we do are we in
30:33a regulated industry? You know, what what are those things that that we do, right? As as an NXT. I think you know, more teams should just to there. I think when you gravitate towards, you know, your pilot experience, we always offer up to customers where we want to help them identify the two or three things that they really want to get out of the product or like meaning like what they want to solve as use cases rather than necessarily trying to just test everything under the sun cuz I don't think that's that really gets you that that true ROI. And I think like the the third point which which should be more obvious is like
31:04which I think is not is like figuring out which vendor is able to prioritize your voice. And I think one of those, you know, one of those obvious things is like maybe using, you know, Harvey as an example. If your in-house team like I think you would think hard about using Harvey necessarily because if you if you're looking to influence the road map because they're 9% focused on on big law firms and that big law experience. Most of the features that they're launching are for that. And so if you have
31:36specific things where you want them to build Slack integrations or things like that, it's likely going to be missed on that you know, on that road map based off priorities. So I think like, you know, focusing on an in-house team or an in-house focused vendor that is more aligned to also maybe the, you know, kind of like that you know, that stage of your company and and being able to really prioritize the things that are important for you. I think that is something that that is really important that I think often gets missed.
32:06Speaker change: So I think vendors are also partly to blame for this because first of all, they're not always very clear on exactly which customer segment they're serving because they might be also exploring themselves or they want to widen the net, let's say, and capture more leads. Yeah, and second of all, vendors all will say that they will bring a curated experience and that they're tailored to the customer's needs. And let's say if a customer actually did the homework, the legal teams did the homework and say we only have these two use cases, what I've been hearing from the legal teams is
32:36that are always trying to upsell and sell them something that they didn't initially ask for. So, if you came for a legal research function, I'm going to try and sell you a contract automation function and a CLM function as well. So, do you agree with that also being the case?
Speaker change: Um you know, surely there's I think vendors uh you know, trying to stretch in in all sorts of directions. Uh I think that's definitely culprit. I think what's uh unique about us is we try to be really honest. Um honestly and and we
33:06we we've identified early on sometimes like hey, looks like you're looking for more CLM with those pre-sanctioned workflows or you know, something from DocuSign or something like that and we actually just make that recommendation up front. We don't try to like you know, force a square peg in a round hole sort of thing. And so, that's something maybe that that is refreshing from folks that chat with us, you know, relative to other to other uh vendors.
Speaker change: For less transparent vendors out there, what do you think is a tactic or question or way for legal teams to be
33:37able to find out whether that vendor is actually a good fit for their core use cases? Cuz their their sales teams are always very slick and they always try to capture every lead. So, if you were in their shoes, what questions would you ask or what would you do to really find that good fit, you know, if the other side isn't as transparent as Rally is?
Speaker change: Yeah, it's great question. Well, one I think use resources like yours, Anna. Uh at Legal Benchmarks, I think that's one great resource. You know, I think I think you know, looking at you know,
34:07some of your the customer success stories is another sort of proxy. Um but really I think getting your hands on the product and but doing that homework in advance where you isolate out sort of those two or three things and get that buy-in from the rest of your team to figure out what are those two or three things that you really want to drill in on and I think you know, you can even ask for more of a even if you don't behind you go into that uh you know, that trial or or that that pilot experience, asking for a more specific tailored demo on those specific use
34:38cases is is maybe a a good approach initially before you even diving deeper with those teams.
Speaker change: Mhm. So, speaking about transparent parency, I'm going to say that again cuz that is a botched. Speaking about transparency, I want to ask about the pricing for Ruly because I've heard from my previous conversations from Preston, which is the CEO of Simple Docs, that most VC-backed companies deliberately hide the pricing for their products because there's pressure from the VCs and also this
35:08understanding that if you hide the pricing, then you can jack up the prices potentially later on. So, what is decision behind Ruly despite being venture-backed listing out the pricing transparently and have you faced any pressure from VCs for you not to do that?
Speaker change: You know, you have to have the right VCs that sort of are in your corner. We haven't had that pressure. You know, our VCs, you know, trust us and and and have us, you know, lead the way. We think it's important to build trust and give
35:38certainty. So, I don't think we've actually changed our pricing in 9 months. Looking back, even we have that We have We want to give that certainty to our customers and and be transparent. And And of course, like we also do volume discounts for over, you know, five users and and ensure that And And so, I think like it's it's it's important. I think it's surprising and we've learned about these pricing games for some of our competitors and they've benefited to our advantage because I
36:08think it's created this confusion in the market where they started with a particular vendor, they were at a certain price, they started to look at us, and then when they went back to that vendor, the price changed or whatnot. And for us, offering that stability, I think has been very welcome.
Speaker change: So, switching gears a bit, if you were to build Ruly from scratch today, what would you do differently?
Speaker change: Originally because of our big tech background, we talked to a lot of lawyers in tech who in Silicon Valley, and I think they were they're in that
36:39early adopter train. And so, that using that as a proxy early on meant that we built even ahead of like where the general market was. So, I think actually talking to more customers again like outside of Silicon Valley in you know, the Midwest in like the South, other markets, right? Um is important to kind of get that pulse especially in the legal market where I think that that greater S curve is actually not the behavior is a little bit different from where
37:09the early adopter train is, which which also tends to be a little bit more you know, DIY and advanced on you know, where AI in general is going.
Speaker change: How do you think a lawyer's work will be different in let's say 20 years time in 2041?
Speaker change: Yeah, I mean that's that's that's a great question. I think I think just you know, starting with what we've seen from last summer to now, you know, just a little under a year. I think one thing that's unique that that unique insight that we've seen is a lot of our customers and and other pilot trials, we
37:41had people you know, we had lawyers kind of going into the the software and trying to run the hardest tasks and and try to show me everything you can do kind of like a really hardcore you know, test drive in a way that's not you know, practical how to use big data. But now we've actually seen a shift where of late we've had the same sort of customers and and new sort of pilot users ask us like, can we curate so there's a little bit less red lines and just make it more practical. We're not
38:12looking for something exhaustive, but we're looking for something that's practical of like what do I actually need to do cuz like what we is more realistic to actually red line with the other side or what are the real takeaways from, you know, some of the legal research. So, we've made those adjustments more recently to actually like dial back some of the the AI results, which was very interesting cuz originally people wanted almost to to test that AI could catch anything. But, now they're like, "Okay, now I I almost have that trust and adoption adoption
38:42curve that's basically happened in the last 9 months." And we've seen that usual behavior shift to now, "Okay, I kind of trust AI." Just really practically what I have to do.
Speaker change: So, I think like we've seen that already happened in a very compressed timeline in these 9 months. So, I think then the question is going out to 2041, that's that's a lot of time, right? If you've seen like based on what's happened in 9 months, I think then you absolutely get to this environment where, you know, in our vision of, you know, the violinist moving to being the orchestra conductor. I think there will be different models
39:12of how you work within the company, within the different teams as well. I think you're going to see something that we circ coin like agent-to-agent workflows. And I think it'll be something where, you know, whatever is the new CRM of the future probably will just talk to the legal agent directly. And then only if there's something that requires a human-to-human involvement, are you going to get involved. But, it's likely, you know, that the coding agent's just going to get, you know, the legal the legal agent involved to look at things around, you know, product counsel issues directly. And then, you
39:44know, the product counsel will get a report on it and everyone kind of moves into this more orchestration layer.
Speaker change: Yeah. So, Paul So, Paul is back into the present moment and reality. And to wrap up, for real this time, this conversation with two quick-fire questions. What's your current test tech stack in terms of like the tools that would be that that have been a game-changer for you? And what resources that you think are something so amazing that people should all check out to stay current on AI?
Speaker change: Yeah. Great great questions. Uh so, for
40:15us, we're pretty uh we we like to keep it pretty light but we have gravitated towards as a as a team uh um either Claude or Gemini. Um part of it is a lot of it's for coding for engineering team but even our designer has has become a vibe coder to kind of bridge the gap between designs and and the front UI level which is very cool. I use Claude a lot for a mix of thought sort of thought partners or developing products. It's actually really great sort of like quick UI testing on sort of things that I have like thesis is in
40:46mind but also I like to use it to kind of mock up things really quickly have a productive conversation with my designer and engineering team. Um a couple of resources for folks to check out is I'm a big uh you know listener of like Lenny's podcast which focuses on a lot of like uh you know product and tech interviews on kind of like you know emerging but also very practical you know takeaways from a lot of it you know in AI and tech. Uh but also like to listen to um Steve Eisman who is one of those guys who is uh big hedge fund investor from
41:18like the big short so one of those guys who shorted you know the market with financial crisis and got that right. Uh but he does a you know his his he's got a podcast out that really talks about the macro investment picture. Um I think that's really important to understand um where the money is funneling into sort of AI or not and how it's impacting SaaS companies and I kind of like to look at that for like the macro picture.
Speaker change: Yeah, thank you so much Brian for sharing your thoughts and the resources and the tools you use. That's a wrap for today and hopefully we'll see more of
41:48really in the upcoming years.
Speaker change: Thank you Anna for featuring us. You know it's been a blast and just an invigorating conversation. Thank you.
Speaker change: Oh wow, I got that on the record. If you like this episode, check out the other LIP deep dives on the Legal Benchmarks channel and subscribe.
