1 00:00:00,480 --> 00:00:01,680 How are you this afternoon? 2 00:00:02,340 --> 00:00:03,150 I'm amazing, Ted. 3 00:00:03,150 --> 00:00:03,870 It's so great to see you. 4 00:00:04,019 --> 00:00:05,190 Yeah, you as well. 5 00:00:05,490 --> 00:00:07,020 We've been trying to do this for. 6 00:00:07,860 --> 00:00:08,820 A while. 7 00:00:09,360 --> 00:00:13,260 Yeah, the schedule has been hectic for all the right reasons, but 8 00:00:13,260 --> 00:00:15,330 I'm really excited that we finally got the chance to sit down. 9 00:00:15,480 --> 00:00:21,119 Yeah, so am I. So am I. It's um, it's a conversation that's overdue, but 10 00:00:21,570 --> 00:00:22,680 I'm really looking forward to it. 11 00:00:22,680 --> 00:00:24,390 We have a really good agenda. 12 00:00:25,080 --> 00:00:27,210 Talk through lots, lots to talk through. 13 00:00:27,790 --> 00:00:31,270 Um, before we jump in, why don't you just do a quick introduction. 14 00:00:31,270 --> 00:00:34,120 I'm sure most people know who you are, but, um, for those that 15 00:00:34,120 --> 00:00:37,270 don't, just a little background on who you are, what you do Yeah. 16 00:00:37,270 --> 00:00:37,930 And where you do it. 17 00:00:38,590 --> 00:00:39,160 Of course. 18 00:00:39,280 --> 00:00:41,755 So, I'm Max, I'm the CO and co-founder of Lara. 19 00:00:42,010 --> 00:00:44,290 Lara is an AI platform for legal work. 20 00:00:44,290 --> 00:00:47,769 And the way that we've really started to think about it is we're building 21 00:00:47,769 --> 00:00:52,210 the place where both humans and agents will be completing legal work together. 22 00:00:52,215 --> 00:00:52,475 And so. 23 00:00:53,385 --> 00:00:57,435 We founded in 2023, just completed our Series D fundraise, where 24 00:00:57,435 --> 00:01:00,105 we raised over $550 million. 25 00:01:00,375 --> 00:01:04,335 Uh, we're about 400 people globally across eight offices, and we're 26 00:01:04,335 --> 00:01:05,325 having the time of our lives. 27 00:01:06,165 --> 00:01:07,065 It's incredible. 28 00:01:07,065 --> 00:01:08,055 It's incredible growth. 29 00:01:08,055 --> 00:01:12,405 Like you and I were on a podcast, I don't know if you remember this like I do. 30 00:01:12,435 --> 00:01:12,825 Yeah. 31 00:01:12,825 --> 00:01:13,245 Like this. 32 00:01:13,245 --> 00:01:16,020 That was, I, you, I, I think you guys were just starting up totally a year ago. 33 00:01:16,020 --> 00:01:16,300 A year and a half. 34 00:01:16,615 --> 00:01:19,255 I think it was more like two, two and a half years ago. 35 00:01:19,255 --> 00:01:23,664 I don't even know if I had my podcast then it was, it was early days. 36 00:01:23,905 --> 00:01:25,015 Well, you had a good setup at least. 37 00:01:25,164 --> 00:01:29,485 Yeah, and it's been it, man, it's been a lot of fun to watch, uh, 38 00:01:29,574 --> 00:01:33,774 you know, watch the whole space grow and watch you guys grow and. 39 00:01:34,330 --> 00:01:38,230 I've seen you've hired a lot of really good people on your team. 40 00:01:38,380 --> 00:01:41,080 Um, we had Kyle Poe one back in December. 41 00:01:41,080 --> 00:01:44,230 Yeah, I, I, I like to think that my job at this point, Ted, is to hire 42 00:01:44,230 --> 00:01:45,760 people much smarter than myself. 43 00:01:45,790 --> 00:01:49,750 And, uh, I've certainly had some pinch me moments where we do a 44 00:01:49,750 --> 00:01:52,630 company-wide town hall and you sort of look out over the group and you 45 00:01:52,630 --> 00:01:57,850 feel like, how on earth did I manage to, um, sort of lure everybody here. 46 00:01:58,180 --> 00:01:59,200 But no, we're grateful. 47 00:01:59,200 --> 00:02:00,010 We're super excited. 48 00:02:00,990 --> 00:02:04,470 Really, I think that's a big part of what, what keeps me going, uh, getting 49 00:02:04,470 --> 00:02:05,880 to work with so many inspiring people. 50 00:02:06,390 --> 00:02:06,780 Yeah. 51 00:02:06,840 --> 00:02:07,230 You know? 52 00:02:07,260 --> 00:02:11,880 Uh, have you ever read the book or heard of the book Good To Great by Jim Collins? 53 00:02:12,600 --> 00:02:13,380 No, but maybe I should. 54 00:02:13,380 --> 00:02:13,770 Yeah. 55 00:02:13,770 --> 00:02:15,540 It's a good, it's one worth reading. 56 00:02:15,570 --> 00:02:21,120 He talks about, uh, he, he goes back and does a study of, I don't know, 57 00:02:21,420 --> 00:02:25,080 hundreds of Fortune 500 companies over really long periods of time. 58 00:02:25,769 --> 00:02:31,200 He finds patterns and anti-patterns for success over the long haul. 59 00:02:31,559 --> 00:02:37,769 And one of the anti-patterns he calls the model of a genius in a thousand 60 00:02:37,769 --> 00:02:44,700 helpers, where, you know, one person, you know, may usually the CEO is, you 61 00:02:44,700 --> 00:02:49,674 know, kind of the, the maestro and the smartest guy in the room and it. 62 00:02:50,475 --> 00:02:56,055 It creates a lot of friction towards scale and um, yeah, it's a great book. 63 00:02:56,205 --> 00:02:59,085 Definitely worth I'll, I'll, I'll pick it up and, uh, that's 64 00:02:59,085 --> 00:03:00,915 certainly not Lara, right? 65 00:03:01,065 --> 00:03:02,655 Yeah, it's not info dash either. 66 00:03:04,425 --> 00:03:09,105 Um, well, why don't we start off with just talking about like the landscape, 67 00:03:09,165 --> 00:03:11,745 which is constantly changing and Yeah. 68 00:03:12,285 --> 00:03:15,975 You know, we had, it's been such a rollercoaster with like the 69 00:03:15,975 --> 00:03:18,015 Claude Legal plugin this year and. 70 00:03:18,720 --> 00:03:23,040 Um, you know, the whole, all this talk of rappers, which. 71 00:03:23,520 --> 00:03:27,540 I'm surprised that nobody's accused us of being a rapper on top of 72 00:03:27,540 --> 00:03:31,320 SharePoint, which, which we are essentially, like, we're built on 73 00:03:31,320 --> 00:03:34,290 top of SharePoint online and Azure. 74 00:03:34,890 --> 00:03:38,490 Um, nobody's called us a rapper yet, so I guess that's a good thing. 75 00:03:39,150 --> 00:03:41,550 Well, I think we're all, we're, we're all wrapping something, right? 76 00:03:42,060 --> 00:03:42,420 Yeah. 77 00:03:42,420 --> 00:03:45,840 You know, I mean, I've heard, I've heard the metaphor. 78 00:03:45,870 --> 00:03:48,510 Well, all apps are just rappers on a database. 79 00:03:49,049 --> 00:03:54,780 And I guess that's true to a certain extent, but, um, how do you, how do you 80 00:03:54,780 --> 00:04:00,269 see the landscape just broadly and where Lara fits in, like, you know, how, what 81 00:04:00,269 --> 00:04:03,060 makes Lara different from its competitors? 82 00:04:03,060 --> 00:04:04,019 Why don't we start with that? 83 00:04:04,829 --> 00:04:05,160 Sure. 84 00:04:05,579 --> 00:04:08,040 Well, look, it's, it's, you know. 85 00:04:08,565 --> 00:04:12,975 End of quarter one of 2026 and the models are where they are. 86 00:04:13,065 --> 00:04:16,965 I think they made a real sort of step change in their capability 87 00:04:16,965 --> 00:04:20,475 and intelligence over Christmas with the Opus 4.5 and 4.6. 88 00:04:21,195 --> 00:04:26,355 And what Lara and where we live in the stack is really sort of the um, central 89 00:04:26,355 --> 00:04:30,975 workplace where a lawyer wakes up in the morning, they open up legum, they 90 00:04:30,975 --> 00:04:33,015 might fire off some work to our agent. 91 00:04:33,525 --> 00:04:35,655 Our agents will start to perform those things. 92 00:04:35,835 --> 00:04:38,680 They will then check in on it, and then we allow the collaboration 93 00:04:38,750 --> 00:04:40,080 both internally with the team. 94 00:04:40,680 --> 00:04:46,080 But also externally with other firms or with their clients and Inc. Increasingly, 95 00:04:46,080 --> 00:04:50,790 we're also seeing large legal departments from um, sort of insurance companies, 96 00:04:50,790 --> 00:04:54,630 banks, large tech companies starting to adopt these tools internally. 97 00:04:55,080 --> 00:05:00,120 So I think there was a lot of initial buzz in 20 24, 20 25 with, 98 00:05:00,210 --> 00:05:01,800 you know, what can this tech do? 99 00:05:02,910 --> 00:05:04,200 Sort to the point. 100 00:05:05,535 --> 00:05:10,815 Scale of adoption is moving fast, and every new tool and new capability 101 00:05:10,815 --> 00:05:15,225 that we add, we can very nicely sort of add to the arsenal of 102 00:05:15,225 --> 00:05:16,575 what our agent can go out and do. 103 00:05:17,655 --> 00:05:22,245 Yeah, it, it really does seem like the world changed in December in terms 104 00:05:22,245 --> 00:05:28,335 of, you know, I think the, to your point, Opus 4.5 and then Opus 4.6, it 105 00:05:28,335 --> 00:05:32,805 really feel, it felt like there was a seismic shift in terms of capability. 106 00:05:34,305 --> 00:05:35,505 Foundation models. 107 00:05:35,895 --> 00:05:40,635 How has that, how has that affected your business? 108 00:05:40,725 --> 00:05:43,515 Um, obviously you're built on top of these models. 109 00:05:43,965 --> 00:05:44,085 Yeah. 110 00:05:44,235 --> 00:05:48,825 But is there a, you know, a mindset of you gotta stay one step ahead 111 00:05:48,825 --> 00:05:50,565 of the sheriff, so to speak. 112 00:05:50,565 --> 00:05:55,575 In terms of building on, on top of that, do they ever overlap in terms of 113 00:05:55,575 --> 00:05:58,395 capabilities that you have in existence? 114 00:05:58,395 --> 00:05:59,050 How do you think about that? 115 00:06:00,135 --> 00:06:03,195 Well look, when we founded LaGuardia in 2023, there was really 116 00:06:03,195 --> 00:06:04,845 just one model to work off of. 117 00:06:05,175 --> 00:06:09,885 It was GPT-3 0.5, uh, which was being hosted on Azure, like you mentioned. 118 00:06:09,885 --> 00:06:14,085 And, um, you know, from that point we've seen models sort of come 119 00:06:14,085 --> 00:06:17,865 and go and every quarter we get new capabilities to work with. 120 00:06:18,405 --> 00:06:23,175 I think one job of, of our team here at Lara is how do we bring the latest and 121 00:06:23,175 --> 00:06:27,435 greatest and how do we sort of squeeze the most juice out of the latest models 122 00:06:27,705 --> 00:06:30,195 for our use cases and for our customers? 123 00:06:30,795 --> 00:06:33,225 Because as you said, Ted, it's moving really fast. 124 00:06:33,315 --> 00:06:37,095 These are large organizations that we work with, large law firms who are 125 00:06:37,095 --> 00:06:41,655 adopting Lara Enterprise wide, and that also means that we need to bring those 126 00:06:41,655 --> 00:06:43,815 capabilities for different practice areas. 127 00:06:44,205 --> 00:06:47,085 So the way that you work with Legian litigation is different from m and 128 00:06:47,085 --> 00:06:48,614 a and it's different from in-house. 129 00:06:48,974 --> 00:06:54,525 And I think the models themselves are a sort of equalizer. 130 00:06:54,825 --> 00:06:56,234 Everybody has access to them, right? 131 00:06:56,265 --> 00:06:59,895 It's sort of the, like the new electricity, but you still need 132 00:06:59,895 --> 00:07:04,664 to put that electricity in a, in a great car, um, or, you know, sort 133 00:07:04,664 --> 00:07:06,494 of get, get the most outta them. 134 00:07:06,854 --> 00:07:09,164 And the way that I like to think about Lago is like a Formula 135 00:07:09,164 --> 00:07:12,015 One car, and then we can hot swap the engine anytime we want. 136 00:07:12,795 --> 00:07:16,905 So the capabilities that the models themselves solve, I 137 00:07:16,905 --> 00:07:17,955 think is ever increasing. 138 00:07:18,405 --> 00:07:20,565 That just allows us to build more on top of them. 139 00:07:21,615 --> 00:07:22,275 Interesting. 140 00:07:22,664 --> 00:07:28,755 And then the, you know, SaaS apocalypse as they call it, um, the Claude 141 00:07:28,755 --> 00:07:33,885 legal plugin, which I have a lot of thoughts on that I've, some of which 142 00:07:33,885 --> 00:07:35,865 I've shared on previous podcasts. 143 00:07:35,865 --> 00:07:41,655 I did a podcast with Ilta recently where we, um, myself, Kat Casey and Dan Zebo. 144 00:07:42,405 --> 00:07:47,805 Um, did a deep dive into what the meaning of it really is. 145 00:07:47,805 --> 00:07:53,895 And, you know, my theory on why there was such a swift market 146 00:07:53,895 --> 00:07:57,735 reaction to really, there were no new capabilities as part of the legal 147 00:07:57,735 --> 00:07:59,775 plugin, like absolutely nothing new. 148 00:08:00,105 --> 00:08:01,635 The plugin architecture existed. 149 00:08:02,145 --> 00:08:07,305 It's some really simple markdown with, you know, six skills that are extremely basic. 150 00:08:07,305 --> 00:08:08,655 There were no new capabilities. 151 00:08:09,015 --> 00:08:14,445 So why did it have such a. A huge impact on the markets. 152 00:08:15,045 --> 00:08:20,145 My theory on that, and I'd be curious to get yours, my theory on that is, 153 00:08:21,525 --> 00:08:28,845 um, the plugin was a release valve for a lot of anxiety in the marketplace 154 00:08:28,905 --> 00:08:37,034 over a lot of capital investment, high valuations, uncertainty about how. 155 00:08:37,455 --> 00:08:42,255 You know how AI is going to impact certain industries, and you saw 156 00:08:42,255 --> 00:08:47,835 companies impacted that shouldn't have been like, you know, TR and Lexus with 157 00:08:47,835 --> 00:08:50,295 all of the proprietary data they have. 158 00:08:50,715 --> 00:08:53,415 LegalZoom got hit hard and that made more sense to me. 159 00:08:53,685 --> 00:08:57,315 I actually see more impact to a LegalZoom business model 160 00:08:57,315 --> 00:08:59,505 than I would to TR or Lexus. 161 00:09:00,135 --> 00:09:02,865 Um, but I, it, it felt like a release valve. 162 00:09:03,355 --> 00:09:07,855 And an excuse for this anxiety to manifest itself. 163 00:09:08,275 --> 00:09:08,545 I don't know. 164 00:09:08,545 --> 00:09:12,235 What is your take on the legal plugin and its impact on the market? 165 00:09:14,155 --> 00:09:22,525 Well, I think you have traditional sauce companies that are more or less 166 00:09:23,155 --> 00:09:29,905 well equipped to execute on what an AI native stack and future will look like. 167 00:09:31,105 --> 00:09:35,280 I think that's independent or sort of, you know, set aside from, 168 00:09:35,730 --> 00:09:39,390 you know, whether, uh, Claude can work with a legal skill or not. 169 00:09:39,390 --> 00:09:39,690 Right. 170 00:09:40,200 --> 00:09:44,460 I think you're somewhat right in that, you know, maybe it was a, a marketable moment 171 00:09:44,460 --> 00:09:48,630 that became an opportunity to, or a reason to sort of, you know, pivot a certain way. 172 00:09:49,650 --> 00:09:54,780 Now I think generally the source that has been built up, like the cost 173 00:09:54,780 --> 00:09:57,240 of writing software is going down. 174 00:09:59,579 --> 00:10:03,270 Opportunity space and the amount of greenfield problems you can go and 175 00:10:03,270 --> 00:10:07,920 solve have actually significantly increased with the model capability. 176 00:10:08,579 --> 00:10:12,209 So we had this really funny moment because we were out racing our series D 177 00:10:12,719 --> 00:10:18,180 at the time of this release, and you had the public, you know, stock sell off, 178 00:10:18,780 --> 00:10:23,880 and we experienced astronomical demand in our own private fundraising round. 179 00:10:24,329 --> 00:10:24,719 And so. 180 00:10:25,770 --> 00:10:31,530 Although there's may, maybe a, a lack of, of trust and excitement around what the 181 00:10:31,589 --> 00:10:35,520 public sales companies can do in terms of executing on this AI native future, 182 00:10:36,120 --> 00:10:40,380 there was a real excitement about, uh, ours and a few, you know, other companies 183 00:10:40,380 --> 00:10:41,699 ability to go and execute on that. 184 00:10:41,699 --> 00:10:45,839 So now I think there's trade-offs, um, and 185 00:10:47,880 --> 00:10:49,020 I am. 186 00:10:50,640 --> 00:10:57,270 Again, sort of very happy that the models keep developing because data is also 187 00:10:57,270 --> 00:11:01,320 playing into the way that our strategy is, which is let's sort of build, you 188 00:11:01,320 --> 00:11:06,420 know, every functionality in Lago like a boat, and when the tide rises, when the 189 00:11:06,420 --> 00:11:11,430 underlying capabilities improve, every single feature that we have gets better. 190 00:11:12,390 --> 00:11:17,280 And I also tried to tell the team, Ted, that we need to build with a very low ego. 191 00:11:18,090 --> 00:11:21,630 You might spend a lot of time building something, you know, nine months ago 192 00:11:21,840 --> 00:11:26,820 that made a lot of sense back then, but it won't make so much sense now that 193 00:11:26,820 --> 00:11:31,170 we're in a world where agents will be executing and doing work on our behalf. 194 00:11:31,680 --> 00:11:31,800 Right. 195 00:11:31,800 --> 00:11:36,420 I think we've moved from the sort of co-pilot, you know, let's iterate 196 00:11:36,420 --> 00:11:40,170 very tightly to, I'm actually gonna let the agent go and work 197 00:11:40,170 --> 00:11:43,080 for a bit and then review the output that it's coming back with. 198 00:11:44,339 --> 00:11:47,699 Yeah, and you know, that was another theory that was put forward in 199 00:11:47,699 --> 00:11:54,780 terms of why the selloff took place, which is with how good agents have 200 00:11:54,780 --> 00:11:59,670 gotten, even automating a ui, right? 201 00:11:59,760 --> 00:12:04,110 That it's going to put pressure on the per seat licensing model. 202 00:12:04,290 --> 00:12:08,040 If you can license an agent to do what three people can do, you 203 00:12:08,040 --> 00:12:10,079 now need one third, the licenses. 204 00:12:10,920 --> 00:12:15,015 And I, you know, I. Like, okay. 205 00:12:15,045 --> 00:12:17,954 But we can adapt to that. 206 00:12:17,954 --> 00:12:22,545 We can, but that's not the only licensing model that's available in the universe. 207 00:12:22,545 --> 00:12:22,785 Right. 208 00:12:22,785 --> 00:12:25,725 We can, that you can get, you can get around that. 209 00:12:26,175 --> 00:12:34,755 Um, how do you see the, you know, agentic phenomenon impacting per user licensing? 210 00:12:34,755 --> 00:12:37,814 'cause I would imagine you're on a per seat license model as well. 211 00:12:37,814 --> 00:12:38,079 Is that accurate? 212 00:12:39,285 --> 00:12:42,675 Yeah, we're on a per seat and, and sometimes consumption as well. 213 00:12:42,735 --> 00:12:48,675 And I think, look, you, you've had, uh, tools for engineers that have been 214 00:12:48,675 --> 00:12:54,165 out for, for a time, if you look at Cursor or sort of the first, um, really 215 00:12:54,165 --> 00:12:58,305 adjunct tools that came out and they've always been consumption based pricing. 216 00:12:58,305 --> 00:13:02,835 And so I think that as a paradigm makes sense for sure. 217 00:13:03,495 --> 00:13:07,615 Um, and then I think there's a question of how do we, as a market, you know. 218 00:13:09,314 --> 00:13:13,995 So that we can control that in an effective way, right? 219 00:13:13,995 --> 00:13:18,735 Like, um, if you're leveraging a lot of AI in an e-discovery process, 220 00:13:19,155 --> 00:13:22,520 you're already used to paying something on a consumption based model. 221 00:13:23,204 --> 00:13:25,935 'cause you could run, you know, hundreds of millions of documents. 222 00:13:26,564 --> 00:13:30,135 And so I think there's a, there is an, an understanding of that. 223 00:13:30,255 --> 00:13:34,574 And as we move more and more into these agent, uh, sort of task 224 00:13:34,574 --> 00:13:38,415 completions, as you explained Ted, you are using more and more tokens. 225 00:13:39,045 --> 00:13:43,485 And so in order for that margin profile and for that, uh, work to actually 226 00:13:43,485 --> 00:13:47,655 make sense, I think there has to be some level of consumption, uh, pricing 227 00:13:47,714 --> 00:13:52,814 associated with it is, is outcome. 228 00:13:53,670 --> 00:14:02,189 Or deliverable based pricing in, in the near term, or is that still further away? 229 00:14:02,939 --> 00:14:07,230 I think that's possible, but I, it's, it's, it's harder because then 230 00:14:07,230 --> 00:14:10,050 you need, you need to build out the taxonomy of the task that you can 231 00:14:10,050 --> 00:14:12,569 solve, right, to a very granular level. 232 00:14:12,660 --> 00:14:17,490 And the different types of tasks that you can solve with leggo, at least 233 00:14:17,490 --> 00:14:20,370 across the different practice areas is really broad and really varied. 234 00:14:22,500 --> 00:14:27,210 We're also seeing firms moving to that from the billable hour, right? 235 00:14:27,210 --> 00:14:30,330 In certain practice areas and in certain geographies. 236 00:14:30,360 --> 00:14:34,920 And so I think all of these developments, the waterfall will sort of happen 237 00:14:35,640 --> 00:14:37,320 slowly, but sort of altogether. 238 00:14:38,730 --> 00:14:44,189 And then I, um, I'm sure you heard of about the, the post on 239 00:14:44,189 --> 00:14:48,870 XI think it was Zach Shapiro who wrote about the Claude Native Firm. 240 00:14:49,560 --> 00:14:55,860 Um, and there's been a lot of back and forth on LinkedIn about, you know, how, 241 00:14:55,920 --> 00:15:03,000 how realistic, um, some of what was put forward actually is in terms of how do 242 00:15:03,000 --> 00:15:05,970 you, how, how do you explain to someone. 243 00:15:06,780 --> 00:15:12,989 What the difference is between leveraging just, you know, 244 00:15:13,140 --> 00:15:17,459 clawed and a bunch of prompts and skills versus leveraging a leg. 245 00:15:18,089 --> 00:15:21,030 How do you differentiate it to people? 246 00:15:21,060 --> 00:15:23,489 How, how, what those different paths look like? 247 00:15:25,074 --> 00:15:29,369 Um, I, I've, I've spent enough time with lawyers to learn to say that it depends, 248 00:15:29,369 --> 00:15:31,890 but it really depends on the problem that you're trying to solve, right? 249 00:15:32,520 --> 00:15:33,750 Let's say you are. 250 00:15:36,750 --> 00:15:38,220 Contracting in an enterprise. 251 00:15:39,390 --> 00:15:43,950 The way that you actually set that up is you want the entire, you know, 252 00:15:43,950 --> 00:15:48,390 company to be able to leverage the same set of playbooks and, and uh, 253 00:15:48,450 --> 00:15:50,010 work, produce by the legal team. 254 00:15:50,760 --> 00:15:53,700 And then you want an agent to go and review all those documents, 255 00:15:53,790 --> 00:15:57,150 create markup, and sort of be ready to send back to the other party. 256 00:15:57,600 --> 00:15:59,220 And you need an example language. 257 00:15:59,220 --> 00:16:01,710 You want different fallbacks for different types of documents. 258 00:16:01,980 --> 00:16:04,510 There's a lot of routing in takeaway that you actually work with that. 259 00:16:06,390 --> 00:16:10,590 Or the model as the engine, again, I think makes a lot of sense. 260 00:16:10,590 --> 00:16:14,310 It's the electricity, but then it's the house or the car that you actually 261 00:16:14,310 --> 00:16:18,030 build around that that becomes valuable for an enterprise to deploy. 262 00:16:18,900 --> 00:16:22,800 And if you look on an m and a transaction, the way that you work with the data 263 00:16:22,800 --> 00:16:25,590 room and the way that you sort of walk through that, the way that you then 264 00:16:25,590 --> 00:16:30,360 take the IP part and the employment part, and the finance or tax part, and 265 00:16:30,360 --> 00:16:33,960 then you bundle all of that together into a single due diligence report. 266 00:16:35,040 --> 00:16:39,300 All the work around where the model is actually working is the stuff that we do. 267 00:16:39,900 --> 00:16:44,400 And then we work really hard to, as, as I said, like get the maximum 268 00:16:44,400 --> 00:16:49,439 juice out of the model so that you can trust that the output that you 269 00:16:49,439 --> 00:16:51,990 get back is based in the context. 270 00:16:52,140 --> 00:16:55,980 With citations, you can market, you can red flag it, you could green 271 00:16:55,980 --> 00:16:59,250 light it, and then everybody can collaborate around that at the same time. 272 00:17:00,030 --> 00:17:03,660 And I think the more time goes on, the bigger and bigger difference. 273 00:17:04,500 --> 00:17:08,940 La Agora has from the foundational models, to be honest with you, how, 274 00:17:09,180 --> 00:17:15,480 how do you think law firms are going to differentiate themselves in this 275 00:17:15,480 --> 00:17:17,640 post, in the post transformation era? 276 00:17:17,970 --> 00:17:25,140 Like historically differentiation, like I actually spun up Claude Cowork and as 277 00:17:25,140 --> 00:17:32,520 Input gave it the AM law 100 list and had it go and list out how firms frame. 278 00:17:33,645 --> 00:17:35,745 Their differentiation, narrative. 279 00:17:36,255 --> 00:17:42,495 And, um, eight out of 100 even mentioned technology, right? 280 00:17:42,615 --> 00:17:43,755 Pretty low percentage. 281 00:17:43,755 --> 00:17:48,195 Historically, how law firms have differentiated themselves is, you know, 282 00:17:48,225 --> 00:17:56,145 their people, their brand, their their wins, their, you know, um, their history. 283 00:17:56,805 --> 00:18:00,405 And as we get into a, taking a tech enabled world. 284 00:18:01,770 --> 00:18:06,150 Technology is going to be part of that differentiation equation, but buying off 285 00:18:06,150 --> 00:18:11,880 the shelf tools isn't differentiation, isn't differentiating by itself. 286 00:18:12,210 --> 00:18:15,900 So how do you think, you know, again, kind of post transformation, 287 00:18:15,960 --> 00:18:19,230 and I don't know when that is gonna be, but we're probably gonna be 288 00:18:19,230 --> 00:18:21,210 transforming for a very long time. 289 00:18:21,630 --> 00:18:25,980 But how are firms going to differentiate themselves from each other? 290 00:18:25,980 --> 00:18:27,300 If everybody's running Lara? 291 00:18:28,350 --> 00:18:30,899 How, how am I different than from the firm down the street? 292 00:18:31,889 --> 00:18:37,110 Well, ever great architect works in cud and ever great designer works 293 00:18:37,110 --> 00:18:42,120 in Figma and every great engineer works in cursor or cloud code. 294 00:18:42,270 --> 00:18:45,294 And so, um, you know, if I get a paint brush. 295 00:18:46,379 --> 00:18:48,179 I could do a pretty shitty self portrait. 296 00:18:48,570 --> 00:18:52,230 If I gave a paintbrush to Michelangelo, it would be an amazing piece of art. 297 00:18:52,830 --> 00:18:56,580 And so the tools is of course, one part of the equation. 298 00:18:56,790 --> 00:19:01,919 Are you equipping your people and your firm with the prerequisites 299 00:19:02,370 --> 00:19:06,300 to go out on the, you know, metaphorical battlefield and win? 300 00:19:06,420 --> 00:19:10,710 You know, you don't wanna show up with a plastic sword and you know, your 301 00:19:10,710 --> 00:19:15,930 opponent has a, has a, you know, metal on, uh, then you're not equipping the 302 00:19:15,930 --> 00:19:17,910 firm with the opportunity to go and win. 303 00:19:18,840 --> 00:19:23,430 And I think la, you know, equivalent, like that's part of, of this stack 304 00:19:23,430 --> 00:19:25,560 required to actually do real work. 305 00:19:26,220 --> 00:19:30,420 Now, the way that you leverage those tools, Ted, that can 306 00:19:30,420 --> 00:19:31,920 be really differentiated. 307 00:19:32,595 --> 00:19:39,045 I know Lara users who are standard deviations ahead on how AI savvy they're, 308 00:19:39,855 --> 00:19:44,955 and the amount of work that they can get done at the quality that they do it with. 309 00:19:44,955 --> 00:19:48,075 The help of systems like Lara is amazing. 310 00:19:48,375 --> 00:19:49,485 It's, it's incredible. 311 00:19:49,545 --> 00:19:52,155 It's like what's watching a magician work. 312 00:19:52,875 --> 00:19:57,225 But then I also know people who, uh, open up La Gora and they give it a 313 00:19:57,225 --> 00:19:59,446 prompt, which is, write me an SPA. 314 00:20:01,274 --> 00:20:05,835 They're unsatisfied with the output and they go, Hey, I can't do my job. 315 00:20:06,345 --> 00:20:07,425 I'm not gonna look at it. 316 00:20:08,325 --> 00:20:10,935 I think that's a very unsuccessful strategy. 317 00:20:11,504 --> 00:20:15,405 And so even if your firm has Lara, if you don't know how to use it, it's 318 00:20:15,405 --> 00:20:16,754 not gonna be very valuable for you. 319 00:20:17,290 --> 00:20:21,794 And so what I think will be a real differentiator is, one, 320 00:20:22,034 --> 00:20:23,205 how you set up the system. 321 00:20:23,835 --> 00:20:27,495 Two, how you enable your people to actually leverage the system. 322 00:20:28,095 --> 00:20:28,965 And then three. 323 00:20:29,820 --> 00:20:36,389 How, um, how much on the frontier you sort of keep pushing and how much do 324 00:20:36,389 --> 00:20:40,740 you think about your processes and how should we redo our m and a process to 325 00:20:40,740 --> 00:20:45,179 actually deliver a quicker turnaround for our client or a better quality work? 326 00:20:45,360 --> 00:20:48,240 Like that's where the differentiation is going to start to seep in 327 00:20:49,379 --> 00:20:51,659 with all of the other things that you also mentioned, right? 328 00:20:51,659 --> 00:20:53,820 The talent, the brand, the people. 329 00:20:54,090 --> 00:20:55,200 This is the whole package. 330 00:20:55,965 --> 00:20:56,445 Yeah. 331 00:20:56,504 --> 00:21:02,445 You know, I, I really think that the firms are going to have to figure out 332 00:21:02,445 --> 00:21:09,104 how to harness and leverage the knowledge and wisdom that led to winning outcomes 333 00:21:09,104 --> 00:21:12,165 for their clients along with ai, right? 334 00:21:12,195 --> 00:21:12,314 Yeah. 335 00:21:12,375 --> 00:21:16,004 That is a lot of times locked up in a very messy. 336 00:21:16,560 --> 00:21:22,320 DMS, um, you know, there are companies out there that are, you know, cro. 337 00:21:22,650 --> 00:21:26,280 Everybody goes, I can never find what I need in the document management system. 338 00:21:26,640 --> 00:21:30,600 And the reason is because there's a terrible signal to noise ratio. 339 00:21:31,080 --> 00:21:31,860 You know, you've got, you. 340 00:21:32,399 --> 00:21:38,760 17 versions of the same document with Finalco Finalco final 341 00:21:39,000 --> 00:21:40,740 on the end of the file name. 342 00:21:41,219 --> 00:21:43,865 It's very difficult to get in there and make sense of you. 343 00:21:43,865 --> 00:21:46,860 You, you know, you've got emails, you've got matters that, 344 00:21:47,010 --> 00:21:49,320 um, never actually transacted. 345 00:21:49,919 --> 00:21:51,665 Uh, but there's still all of that. 346 00:21:52,304 --> 00:21:57,254 Um, all of those documents in the DMS that if you just turn AI loose 347 00:21:57,585 --> 00:22:01,875 on a DMS, it's going to have the same struggles that a person is that's 348 00:22:01,875 --> 00:22:04,845 trying to find something, um, useful. 349 00:22:04,845 --> 00:22:07,335 That's actually, that actually created a winning outcome. 350 00:22:07,754 --> 00:22:13,754 So how do you think about how law firms are going to leverage that content 351 00:22:13,754 --> 00:22:19,064 and wisdom and the things that created winning outcomes in the AI future? 352 00:22:19,095 --> 00:22:20,235 What does that look like in your mind? 353 00:22:21,495 --> 00:22:25,005 Well, I think that's one part of the puzzle, but, but certainly not all of it. 354 00:22:25,395 --> 00:22:33,105 Um, and what I'd almost, you know, sort of rephrase it to is how do you provide an 355 00:22:33,105 --> 00:22:39,885 agent or your AI software with the context and the capability to fetch that context 356 00:22:40,395 --> 00:22:42,315 so that it can do real work for you? 357 00:22:43,065 --> 00:22:43,305 Right. 358 00:22:43,455 --> 00:22:44,235 And that context. 359 00:22:45,345 --> 00:22:50,205 In your dms, it can live in your email, it could live online, could live in a, 360 00:22:50,535 --> 00:22:53,145 you know, legal research, uh, repository. 361 00:22:54,135 --> 00:22:54,555 And 362 00:22:57,135 --> 00:23:04,575 as you say, navigating these systems by hand as a human is really inefficient and. 363 00:23:06,225 --> 00:23:10,965 Luckily for US, AI and software is a lot faster, and so you can actually let the 364 00:23:10,965 --> 00:23:17,625 ai, you know, query against the databases using either APIs or MZP, and that starts 365 00:23:17,625 --> 00:23:23,385 to get, you know, starts to create the ability for the agents to actually go 366 00:23:23,385 --> 00:23:24,705 and fetch the context that they need. 367 00:23:25,455 --> 00:23:29,024 But what you also then want is you want the orchestration engine and you 368 00:23:29,024 --> 00:23:30,524 want the layer on top that actually. 369 00:23:31,485 --> 00:23:35,354 It tells the agent, this is what you need to do in what order. 370 00:23:35,715 --> 00:23:38,715 This is your plan and this is your execution in order to actually 371 00:23:38,715 --> 00:23:40,215 produce the output that we want here. 372 00:23:40,574 --> 00:23:43,665 Whether that's fetching precedent or going out online to find something. 373 00:23:44,145 --> 00:23:47,385 So again, I think this is sort of, you know, the modern stack here 374 00:23:47,504 --> 00:23:48,705 is really starting to develop. 375 00:23:49,455 --> 00:23:50,054 How about. 376 00:23:50,385 --> 00:23:56,595 Like org structure and, and, um, how big a factor that is with success. 377 00:23:56,955 --> 00:24:00,765 So, for example, you know, I've been selling into the legal market for 20 378 00:24:00,765 --> 00:24:06,435 years and I, historically they have, from my perspective, under invested in 379 00:24:06,435 --> 00:24:11,445 technology and many times in knowledge management in general, knowledge 380 00:24:11,445 --> 00:24:12,975 management has become a catchall. 381 00:24:13,425 --> 00:24:13,995 It's like. 382 00:24:14,745 --> 00:24:19,965 You know, they do everything from actually true knowledge management or we call it 383 00:24:19,965 --> 00:24:25,544 Cam Classic, which is, you know, building and maintaining precedent libraries and, 384 00:24:25,875 --> 00:24:27,945 you know, base, base and model documents. 385 00:24:28,395 --> 00:24:34,004 Um, but they also have really been thrown into the fray to just become. 386 00:24:34,649 --> 00:24:36,450 Change management for, for it. 387 00:24:36,629 --> 00:24:40,050 We've got a new DMS coming and you know, you're gonna help us 388 00:24:40,050 --> 00:24:41,280 figure out how to organize it. 389 00:24:41,850 --> 00:24:44,850 Like what, what do you think about new roles in terms of, we're 390 00:24:44,850 --> 00:24:49,590 hearing all of this Palantir forward deployed engineer has become all 391 00:24:49,590 --> 00:24:51,810 the rage and, and legal engineer. 392 00:24:52,230 --> 00:24:56,220 Like from your perspective, the, the, the firms that have been most 393 00:24:56,220 --> 00:24:57,840 successful deploying your product. 394 00:24:58,290 --> 00:25:00,750 What sort of, what have they done differently from an 395 00:25:00,750 --> 00:25:01,950 organizational perspective? 396 00:25:02,429 --> 00:25:02,639 Yeah. 397 00:25:03,210 --> 00:25:08,610 Well, we were early, uh, to, to, to coin the term forward deployed legal engineer. 398 00:25:08,610 --> 00:25:12,480 And I think it's certainly a way that we've been able to work tightly with 399 00:25:12,480 --> 00:25:16,380 our clients from the very beginning to both co-create the product and 400 00:25:16,380 --> 00:25:18,480 make sure that we drive the adoption. 401 00:25:18,480 --> 00:25:20,790 That then leads to the ROI for these teams. 402 00:25:21,510 --> 00:25:21,570 Um. 403 00:25:23,250 --> 00:25:28,655 As you say, like we've seen every possible permutation of organizational design and, 404 00:25:28,660 --> 00:25:32,850 and team staffing from, from the firms and organizations that we work with. 405 00:25:32,850 --> 00:25:39,720 But generally, um, I think if you create the team whose role and 406 00:25:39,720 --> 00:25:44,760 responsibility is it is to drive innovation, you then need to give them 407 00:25:44,880 --> 00:25:48,150 enough political power and sort of, um. 408 00:25:51,975 --> 00:25:54,675 Opportunity to actually go and drive that change. 409 00:25:54,825 --> 00:25:59,715 Um, the most unsuccessful things we've seen is when you invest a ton 410 00:25:59,715 --> 00:26:03,284 of time in building up an amazing team, and then, you know, you 411 00:26:03,405 --> 00:26:07,215 just don't let them work directly with partners or drive the change. 412 00:26:07,875 --> 00:26:13,185 And, and then I think also most people underestimate the. 413 00:26:14,220 --> 00:26:20,370 Change that is required on an individual's sort of work level in order to 414 00:26:20,460 --> 00:26:23,040 properly get the maximum efficiency. 415 00:26:23,910 --> 00:26:27,840 And if you get a sort of one-on-one, you know, you start to learn how to 416 00:26:27,840 --> 00:26:29,850 work a bit with an assistant like. 417 00:26:30,185 --> 00:26:32,465 That's great, but that's sort of the baseline. 418 00:26:32,645 --> 00:26:37,445 Um, you need to take serious time out of your day to rethink how 419 00:26:37,445 --> 00:26:39,365 you're going to approach your work. 420 00:26:39,754 --> 00:26:45,004 Now, given the tools at your disposal and some change management teams and 421 00:26:45,004 --> 00:26:50,284 innovation teams and M teams, I think have been very effective in helping partners 422 00:26:50,284 --> 00:26:52,450 and practices navigate those changes. 423 00:26:53,165 --> 00:26:56,225 And they almost approach it as process engineers in a way. 424 00:26:56,375 --> 00:26:56,465 Mm-hmm. 425 00:26:56,705 --> 00:26:59,314 So they sort of map out like, okay, this is what we're doing manually. 426 00:26:59,685 --> 00:27:04,605 Now with this new tool, how could we approach it differently instead of going, 427 00:27:04,935 --> 00:27:08,745 okay, for all of these different, you know, here I might use a little bit of 428 00:27:08,745 --> 00:27:10,305 AI here, I might use a little bit of ai. 429 00:27:10,605 --> 00:27:12,165 They sort of approach it a bit more holistically. 430 00:27:12,764 --> 00:27:18,795 Yeah, and you know, historically legal work has done largely on a bespoke basis. 431 00:27:19,335 --> 00:27:20,475 You could go to the same. 432 00:27:21,659 --> 00:27:26,790 Law firm in the same practice and ask for a commercial lease within 433 00:27:26,790 --> 00:27:30,450 the real estate practice at the same law firm and get back four different. 434 00:27:31,230 --> 00:27:33,270 Results from four different attorneys, right? 435 00:27:33,540 --> 00:27:37,500 Because there's, do you know, everybody has their and, and that has been a result 436 00:27:37,530 --> 00:27:42,600 of lessons that these lawyers have learned along the way that they've incorporated 437 00:27:42,600 --> 00:27:47,580 these learnings into their own individual kind of bespoke work processes. 438 00:27:47,940 --> 00:27:51,570 And now we're entering into an era where we're becoming tech enabled. 439 00:27:51,990 --> 00:27:56,520 And this bespoke model is at odds with that if we're gonna scale. 440 00:27:57,360 --> 00:28:04,110 So how do you see, um, it seems like a herding of cats, uh, if you 441 00:28:04,110 --> 00:28:08,550 will, getting lawyers to standardize on their processes again, even 442 00:28:08,550 --> 00:28:10,260 within the same firms sometimes. 443 00:28:10,260 --> 00:28:12,064 So how have you seen that play out? 444 00:28:13,715 --> 00:28:14,804 Well, I 445 00:28:18,540 --> 00:28:21,689 think there's finally tools and opportunities that 446 00:28:21,689 --> 00:28:24,840 motivate doing that because. 447 00:28:25,830 --> 00:28:30,930 It's not just a linear improvement on a way you used to do something. 448 00:28:31,290 --> 00:28:37,020 This is a complete game changer in terms of how you're delivering your work, right? 449 00:28:37,260 --> 00:28:41,280 And it's like going from, you know, pen and paper to the computer. 450 00:28:41,280 --> 00:28:43,680 Like it's a big shift in terms of how you actually work. 451 00:28:43,740 --> 00:28:48,690 And I think the reality is that, um, you have real champions. 452 00:28:52,395 --> 00:28:57,315 Sort of really adopt early and show everybody else what's possible. 453 00:28:58,034 --> 00:29:01,665 And that needs to happen internally in order for the rest 454 00:29:01,665 --> 00:29:02,955 of the organization to follow. 455 00:29:03,554 --> 00:29:07,034 And so one thing that we often do when we're part of our sort of broader 456 00:29:07,034 --> 00:29:11,054 rollouts, but even in pilots, is that we really work with the firm 457 00:29:11,054 --> 00:29:13,185 to identify who this sort of, um. 458 00:29:15,045 --> 00:29:19,575 Fire carriers are gonna be because those are the people that we need to lift up 459 00:29:19,575 --> 00:29:23,835 internally and that we need to highlight because they're gonna be able to drag 460 00:29:23,835 --> 00:29:25,065 the rest of the organization with them. 461 00:29:25,695 --> 00:29:31,275 And I think increasingly Ted, there's also an expectation of their clients to 462 00:29:31,275 --> 00:29:35,925 start to leverage these tools to drive better value and better outcome for them. 463 00:29:36,405 --> 00:29:40,845 Because I actually don't no longer think that, that it's an excuse that, 464 00:29:40,850 --> 00:29:44,205 oh, oh, we haven't started looking at this, or, or something like that. 465 00:29:44,595 --> 00:29:51,885 Now, I think it's very hard to actually motivate not using AI for certain things. 466 00:29:52,425 --> 00:29:54,435 And even if you do things by hand. 467 00:29:55,139 --> 00:29:59,459 You should certainly do it with AI as well, because you might find 468 00:29:59,459 --> 00:30:00,570 things that you otherwise wouldn't. 469 00:30:00,990 --> 00:30:03,179 And so, or find strategies, right? 470 00:30:03,179 --> 00:30:07,530 Like on a, on a litigation case, you might even spend more time if you 471 00:30:07,530 --> 00:30:11,879 work with AI because you're gonna find out more creative strategies 472 00:30:12,149 --> 00:30:13,770 that you can potentially follow. 473 00:30:14,280 --> 00:30:14,669 And so. 474 00:30:16,110 --> 00:30:21,420 I think that there's gonna be sort of multiple classes of, of, of, of, um, 475 00:30:21,450 --> 00:30:23,100 sophistication that comes out here. 476 00:30:23,100 --> 00:30:28,470 But I'm certain that the, you know, really AI empowered and AI native 477 00:30:28,710 --> 00:30:30,000 lawyers will do extraordinarily well. 478 00:30:31,550 --> 00:30:32,840 Yeah, that makes a lot of sense. 479 00:30:32,900 --> 00:30:36,830 What do you think about the chief AI officer role at a law firm? 480 00:30:36,890 --> 00:30:39,380 You know, historically we've had chief innovation officers. 481 00:30:39,950 --> 00:30:46,730 Um, chief AI officer feels a little more narrow than, than just innovation. 482 00:30:46,850 --> 00:30:46,910 Yeah. 483 00:30:46,910 --> 00:30:51,560 I mean, look, I think AI is the biggest existential opportunity and 484 00:30:51,740 --> 00:30:56,150 maybe threats that's happened to the industry in a very, very long time. 485 00:30:56,480 --> 00:30:56,810 Um. 486 00:30:57,825 --> 00:31:04,004 I'm frankly surprised by how little resources many firms are allocating 487 00:31:04,215 --> 00:31:08,865 to this, um, which is understandable given the way that the, you know, 488 00:31:09,045 --> 00:31:10,754 profit sort of gets split every year. 489 00:31:11,475 --> 00:31:15,045 But then again, I think this is just something you have to get right, 490 00:31:15,615 --> 00:31:18,945 and not only do you have to get it right, if you're in a very competitive 491 00:31:18,945 --> 00:31:23,415 environment, you need to get it right more so than your competitors. 492 00:31:24,165 --> 00:31:28,695 And I think that requires real resources, real time, real energy and commitment 493 00:31:28,695 --> 00:31:32,895 from both your partner and the firm. 494 00:31:33,525 --> 00:31:34,755 And you have to work together on that. 495 00:31:35,355 --> 00:31:39,225 I think we've, you know, certainly felt that. 496 00:31:40,380 --> 00:31:43,830 In the biggest deployments that we've done and the tightest partnerships 497 00:31:43,830 --> 00:31:49,440 we have, we have a very sort of free flowing, you know, sharing of information 498 00:31:49,440 --> 00:31:51,150 and what works, what doesn't work. 499 00:31:51,420 --> 00:31:54,060 But then of course, that stays within that specific firm. 500 00:31:54,060 --> 00:31:58,080 And so I think we're seeing different strategies being applied, and I think 501 00:31:58,770 --> 00:32:03,570 the chief AI officer could be, could, could work or not work, probably most 502 00:32:03,570 --> 00:32:06,960 dependent on the person and how much, um, you know, runway you actually 503 00:32:06,960 --> 00:32:08,880 give them how empowered they are. 504 00:32:08,970 --> 00:32:10,080 I think so. 505 00:32:10,170 --> 00:32:10,530 Yeah. 506 00:32:11,280 --> 00:32:12,240 Agreed. 507 00:32:12,660 --> 00:32:17,520 Um, you touched on something that I wanna dive into a little bit, and 508 00:32:17,520 --> 00:32:24,720 that is investment and ROI, um, the law firms, at least here in the US 509 00:32:24,720 --> 00:32:29,850 are primarily partnerships and they mostly operate on a cash basis. 510 00:32:30,480 --> 00:32:30,660 Mm-hmm. 511 00:32:31,050 --> 00:32:34,380 Um, which means that they clear the books at the end of the year, they distribute. 512 00:32:35,055 --> 00:32:40,335 Profits to the partners and they start the year at a zero and do it all over again. 513 00:32:41,145 --> 00:32:46,335 And we are entering into an era where there's gonna need to be capital 514 00:32:46,335 --> 00:32:52,005 investment, short, mid, and long-term capital investment to build out the, 515 00:32:52,035 --> 00:32:56,895 the AI infrastructure because it's gonna be more than just buying lag seats. 516 00:32:56,925 --> 00:32:57,255 Right. 517 00:32:57,255 --> 00:32:58,035 There's, there's. 518 00:32:59,054 --> 00:33:04,004 Um, there is foundational level work at an infrastructure level that needs to happen 519 00:33:04,004 --> 00:33:07,995 to support, um, what AI runs on top of. 520 00:33:08,685 --> 00:33:15,945 So, you know, I've been beating the drum that law firms have historically 521 00:33:16,004 --> 00:33:21,735 over indexed on ROI, and I've been an advocate of quit thinking about 522 00:33:21,735 --> 00:33:25,665 ROI right now because it's gonna be hard to calculate a number. 523 00:33:26,280 --> 00:33:29,969 A spreadsheet because we're, we're in, we're in the midst of change, right? 524 00:33:29,969 --> 00:33:32,669 Like it actually, you may have a negative ROI because you're gonna 525 00:33:32,669 --> 00:33:34,469 reduce billable hours for right now. 526 00:33:34,709 --> 00:33:40,290 But you're learning the, the, the, the r the return in ROI is your learning. 527 00:33:41,010 --> 00:33:45,510 Um, and this is where, this is an era of r and d. Um. 528 00:33:46,560 --> 00:33:50,429 You know, long term there has to be ROI or you can't continue to invest. 529 00:33:50,730 --> 00:33:55,050 But in the short term, I feel like law firms are, are overindexing, not 530 00:33:55,050 --> 00:34:00,719 all of them, but a good percentage are overindexing on what is my ROI right now. 531 00:34:01,679 --> 00:34:02,490 Do you feel that way? 532 00:34:02,490 --> 00:34:07,350 I mean, one, one very clear ROI that we've identified across many of the, of the 533 00:34:07,439 --> 00:34:12,330 firms that we work with is the reduction of work where you're actually not billing. 534 00:34:13,905 --> 00:34:17,325 Which then creates the opportunity to go out and either pitch for more work 535 00:34:17,835 --> 00:34:19,875 or, you know, actually bill for more. 536 00:34:20,385 --> 00:34:23,505 And many of the firms that we work with are, they're at, they're 537 00:34:23,505 --> 00:34:25,545 sort of at the supply constraint. 538 00:34:25,545 --> 00:34:29,295 Like they actually cannot take on more work because they're so fully staffed. 539 00:34:29,925 --> 00:34:36,735 And when, you know, I agree with you that regardless of how you 540 00:34:36,735 --> 00:34:38,175 want to calculate a number, like. 541 00:34:38,730 --> 00:34:42,000 It's kind of go time, like it's here, it's here to stay. 542 00:34:42,270 --> 00:34:43,290 It's undeniable. 543 00:34:43,920 --> 00:34:50,100 However, I, I also do appreciate, you know, the, the ability, the ability 544 00:34:50,100 --> 00:34:51,300 to build a business case around it. 545 00:34:52,110 --> 00:34:56,280 And I think that can come either in the form of sort of, you know, aggregated 546 00:34:56,280 --> 00:35:02,430 numbers or data, but also just the fact that you were able to win that deal. 547 00:35:03,150 --> 00:35:09,090 Because you could tell a story of how your firm leverages AI and how you as a partner 548 00:35:09,090 --> 00:35:14,190 and as a team will leverage AI to get them the best outcome at the best price, right? 549 00:35:14,940 --> 00:35:19,860 And so I think it's no longer a question of do we use this to be faster? 550 00:35:19,860 --> 00:35:21,150 Will we reduce billable hours? 551 00:35:21,150 --> 00:35:26,160 Like if your top firm, you have to use AI to deliver the best 552 00:35:26,160 --> 00:35:30,150 services, the highest quality services at the quickest turnaround. 553 00:35:31,425 --> 00:35:37,935 So there have been studies from McKinsey, Morgan Stanley, Goldman 554 00:35:37,935 --> 00:35:44,625 Sachs, that are estimating 30 to 50% of associate hours being automated away. 555 00:35:45,375 --> 00:35:47,685 And I we're, we're, we're not quite there yet. 556 00:35:47,925 --> 00:35:51,225 But, um, I think we're heading in that, in that direction. 557 00:35:51,795 --> 00:35:52,480 I'm curious about. 558 00:35:53,625 --> 00:35:58,485 With this trajectory that we're on that seems to be moving steeply, 559 00:35:58,695 --> 00:36:00,045 uh, upward into the right. 560 00:36:00,465 --> 00:36:07,515 How long do we have before we're in a place where 30 to 50% of associate 561 00:36:07,515 --> 00:36:12,855 hours are no longer part of the revenue stream for these law firms? 562 00:36:13,215 --> 00:36:17,385 Well, one thing those studies don't take into account is how 563 00:36:17,385 --> 00:36:20,085 much more work there's going to be. 564 00:36:21,285 --> 00:36:24,795 So that's looking at the bucket of work that's available today. 565 00:36:25,395 --> 00:36:29,595 Which part of it do we think that AI can do a, you know, pretty good job of out of 566 00:36:29,595 --> 00:36:35,115 the box or with a system like ro And what they fail to do is they fail to say, okay, 567 00:36:35,595 --> 00:36:38,925 well, because agents will be doing work, 568 00:36:41,805 --> 00:36:43,665 companies will sue more other companies. 569 00:36:44,370 --> 00:36:48,720 There will be more deals, there will be more legal diligence 'cause you're 570 00:36:48,720 --> 00:36:50,160 gonna want to do it from the sell side. 571 00:36:50,160 --> 00:36:52,650 And there's more companies on the buy side who are gonna want to 572 00:36:52,650 --> 00:36:57,990 diligence the company 'cause it's gonna be faster and more accessible. 573 00:36:58,920 --> 00:37:04,170 And so even if there's less associates hours, doing particular 574 00:37:04,170 --> 00:37:08,940 tasks like document review and a diligence, doesn't mean that we'll 575 00:37:08,940 --> 00:37:10,080 necessarily have fewer associates. 576 00:37:11,055 --> 00:37:11,924 Will be more work. 577 00:37:12,315 --> 00:37:16,634 Now, how that gets distributed, where in the value chain, I think is a little 578 00:37:16,634 --> 00:37:22,694 bit sort of TBD, but I think that the associates who are coming out of, of 579 00:37:23,595 --> 00:37:29,565 law school with the skills to leverage AI and who are business-minded and 580 00:37:29,565 --> 00:37:33,075 problem solver and charismatic and can work on clients, sort of, you 581 00:37:33,075 --> 00:37:36,165 know, understand that relationship, those will do extraordinarily well. 582 00:37:36,735 --> 00:37:40,185 And it's also one of the reasons that why we're really excited about our law school 583 00:37:40,185 --> 00:37:41,895 program, which we're launching this week. 584 00:37:42,315 --> 00:37:46,605 So we've partnered up with nine of the leading law schools in the US and not 585 00:37:46,605 --> 00:37:49,694 only are we sort of, you know, getting people access to the system, we're 586 00:37:49,694 --> 00:37:55,095 working with the deans and the law school professors to construct curriculums 587 00:37:55,424 --> 00:38:01,065 that will equip the law school students with the skills to actually bring this 588 00:38:01,065 --> 00:38:02,835 into their workplace once they graduate. 589 00:38:02,835 --> 00:38:07,815 And so we're trying to take responsibility for the full, you 590 00:38:07,815 --> 00:38:09,375 know, in Swedish we would say from. 591 00:38:11,025 --> 00:38:15,254 From sort of corn to bread, but sort of all the way across the value chain. 592 00:38:16,424 --> 00:38:16,995 Interesting. 593 00:38:17,444 --> 00:38:18,345 What are 594 00:38:20,595 --> 00:38:25,365 three things that AI is really good at in legal work, and what are three, three 595 00:38:25,665 --> 00:38:28,395 things that it's not good at today? 596 00:38:30,915 --> 00:38:32,805 AI is really good at reading. 597 00:38:33,105 --> 00:38:37,035 It's really good at finding things in documents, contracts, 598 00:38:37,335 --> 00:38:38,085 whatever it might be. 599 00:38:38,640 --> 00:38:40,379 That it's extraction is really good. 600 00:38:41,940 --> 00:38:46,440 It's really good at looking at a lot of data, big buckets of data 601 00:38:46,529 --> 00:38:52,049 and finding the right thing, sort of needled in a haystack, a bit dependent 602 00:38:52,049 --> 00:38:53,279 on how you structure your data. 603 00:38:53,399 --> 00:38:55,290 But AI is really good at that. 604 00:38:56,399 --> 00:38:58,740 I think AI is really good at ideation. 605 00:38:59,040 --> 00:39:04,560 He's actually really good at, um, finding ideas and seeing patterns that 606 00:39:04,560 --> 00:39:07,589 aren't necessarily as natural to humans. 607 00:39:08,310 --> 00:39:13,830 So ideating on a case strategy or on how you might be able to rephrase a paragraph 608 00:39:13,830 --> 00:39:16,410 or a clause to actually get that through. 609 00:39:17,160 --> 00:39:23,759 Extraordinarily good at what AI maybe isn't as good as today is. 610 00:39:25,800 --> 00:39:33,000 When the context overflows and you, for instance, need to understand how a 611 00:39:33,000 --> 00:39:39,060 very complicated legal research query, um, should be answered using a ton of 612 00:39:39,060 --> 00:39:44,400 different, um, you know, case law and precedent, uh, it's difficult for the 613 00:39:44,400 --> 00:39:49,230 agents to actually look at all of that and then come up with a definitive answer 614 00:39:49,680 --> 00:39:52,320 because they don't necessarily understand how everything sort of ties together. 615 00:39:54,620 --> 00:40:00,375 I think AI is not very good at very complex drafting, right? 616 00:40:00,375 --> 00:40:05,384 Like if you needed to draft a full SPA from scratch, it'll not do a good job. 617 00:40:05,835 --> 00:40:10,905 However, if you provide it a precedent in context, it does a really good job. 618 00:40:11,595 --> 00:40:13,965 So again, it sort of goes back to is AI good out of the box? 619 00:40:14,400 --> 00:40:18,600 Or is AI good within an ecosystem like leg where it can actually perform that 620 00:40:18,600 --> 00:40:22,680 task Well, so I think the future of software is a very integrated one. 621 00:40:23,190 --> 00:40:28,290 It's where you are able to launch agents, give them tasks, give them 622 00:40:28,290 --> 00:40:32,790 commands, and they're going to be able to work with the rest of your ecosystem 623 00:40:32,790 --> 00:40:34,710 and your stack in order to do that. 624 00:40:35,460 --> 00:40:40,529 And so for legal research specifically, one of the reasons why, um, I think that 625 00:40:40,680 --> 00:40:43,440 task has been quite hard is because. 626 00:40:45,780 --> 00:40:51,210 AI always gave the correct answer on research query that 627 00:40:51,210 --> 00:40:52,890 would sort be profoundly. 628 00:40:53,685 --> 00:40:57,645 Scary thing like that means that it's always correct no matter what. 629 00:40:58,155 --> 00:41:01,725 And the reason why it's hard to make it correct all the time for every type 630 00:41:01,725 --> 00:41:06,435 of query is that the problem space is so large because you can ask so 631 00:41:06,435 --> 00:41:07,635 many different types of questions. 632 00:41:08,235 --> 00:41:11,355 And what's important is, of course, one, the underlying data that you 633 00:41:11,355 --> 00:41:14,715 work with, but two, how the data is structured and how you actually 634 00:41:14,715 --> 00:41:16,135 let an agent traverse that data. 635 00:41:17,130 --> 00:41:18,210 And now how you make sense of it. 636 00:41:18,330 --> 00:41:21,810 You know, you need a Tator, you need everything to sort of line up to this. 637 00:41:22,470 --> 00:41:26,280 And we've invested a lot of resources in our US legal research. 638 00:41:26,430 --> 00:41:31,320 And so with both state and federal case law, um, we're now starting to 639 00:41:31,320 --> 00:41:35,730 work more and more with how all of this can be tied together in a very 640 00:41:35,730 --> 00:41:41,310 cohesive, um, experience, both for the user and for the agents that are 641 00:41:41,310 --> 00:41:42,600 then leveraging all of this together. 642 00:41:43,620 --> 00:41:44,400 That makes sense. 643 00:41:44,880 --> 00:41:45,960 So, um. 644 00:41:46,605 --> 00:41:50,205 We only have a few minutes left, but one area I wanted to touch on 645 00:41:50,205 --> 00:41:54,225 that I know it's, IM, uh, important to you is like collaboration. 646 00:41:54,225 --> 00:41:54,285 Yeah. 647 00:41:54,765 --> 00:41:58,185 So Info Dash is a legal collaboration platform. 648 00:41:58,605 --> 00:41:58,964 Yes. 649 00:41:58,964 --> 00:42:05,205 We do not compete with Lara, um, or Harvey or any other practice of law tool. 650 00:42:05,205 --> 00:42:05,805 We really. 651 00:42:06,900 --> 00:42:11,850 Um, we are operate in the compliment, like I think the practice of law. 652 00:42:11,880 --> 00:42:17,430 If you were to draw a pie chart of all of the use cases that fall into 653 00:42:17,940 --> 00:42:21,270 the collaboration bucket, how law firm clients want to collaborate 654 00:42:21,270 --> 00:42:24,360 with their law firms, um, the. 655 00:42:25,995 --> 00:42:29,325 The practice of law represents a fairly narrow slice today. 656 00:42:29,535 --> 00:42:32,745 I think that slice will grow wider as we become tech enabled. 657 00:42:33,404 --> 00:42:39,345 But there are things like, you know, um, matter intelligence, so legal 658 00:42:39,345 --> 00:42:45,254 team information, um, you know, legal project management tasks, uh, 659 00:42:45,285 --> 00:42:47,895 financial information reporting. 660 00:42:47,984 --> 00:42:48,734 Um. 661 00:42:49,904 --> 00:42:52,904 You know, knowledge management, there's all these other things outside. 662 00:42:53,145 --> 00:42:57,345 You guys announced your portal product like two days before TLTF summit 663 00:42:57,765 --> 00:42:58,904 and everybody's coming up to me. 664 00:42:58,904 --> 00:42:59,955 They're like, are you worried about la? 665 00:43:00,525 --> 00:43:05,174 I'm like, you clearly don't understand what a legal extranet platform does. 666 00:43:05,505 --> 00:43:10,725 Yes, law firms will want to collaborate on work product with their customers, 667 00:43:10,725 --> 00:43:11,700 and I think that's where you guys. 668 00:43:12,645 --> 00:43:15,524 Play, but there's so, so much more to that. 669 00:43:15,524 --> 00:43:19,785 But tell me your, like long-term vision about how you see Lara 670 00:43:19,785 --> 00:43:21,464 playing in that collaboration space. 671 00:43:22,484 --> 00:43:26,205 Yeah, so I think it's, it's really sort of two things at the same time, um, we're 672 00:43:26,205 --> 00:43:31,634 working with firms who are looking for ways to scale their delivery method. 673 00:43:32,325 --> 00:43:33,944 Take Debevoise as an example. 674 00:43:34,545 --> 00:43:38,205 Debevoise and the partner Avi there just launched their star 675 00:43:38,205 --> 00:43:41,399 portal on Laua and they're working with clients like Blackstone. 676 00:43:42,630 --> 00:43:45,330 They're also working with other private equity firms, and they're 677 00:43:45,330 --> 00:43:50,490 allowing all those private equity firms to access structured workflows, 678 00:43:50,550 --> 00:43:57,300 work products, databases, and content that gets them value faster. 679 00:43:58,110 --> 00:44:00,330 And so what they're, what they're doing is they're, they're 680 00:44:00,360 --> 00:44:01,920 scaling their delivery method. 681 00:44:03,150 --> 00:44:06,210 And I think that's very interesting because that's a way to actually 682 00:44:06,210 --> 00:44:11,130 move away from, you know, how you're actually leveraging the billable hour. 683 00:44:11,910 --> 00:44:17,640 And I think the real pioneers and the client excitement is real. 684 00:44:18,209 --> 00:44:25,350 They really value the progressiveness of this. 685 00:44:26,399 --> 00:44:29,399 On the other hand, you have, um, enterprises. 686 00:44:31,379 --> 00:44:34,770 That are working with many firms who want to consolidate the 687 00:44:34,770 --> 00:44:35,879 way that that actually works. 688 00:44:35,970 --> 00:44:40,230 So instead of having, you know, 20 panel firms and the way that you 689 00:44:40,230 --> 00:44:44,460 manage all of that work, um, they want to be able to work jointly 690 00:44:44,730 --> 00:44:46,470 on specific matters with them. 691 00:44:47,160 --> 00:44:47,580 And so. 692 00:44:48,750 --> 00:44:51,900 One of the things that we experienced as part of our fundraise 693 00:44:51,960 --> 00:44:55,290 is that collaborating with our law firm is quite inefficient. 694 00:44:55,800 --> 00:45:01,710 It's all on email and when we have, um, to-dos or when we need to communicate 695 00:45:01,710 --> 00:45:06,030 or when I need to go and read our shareholders agreement, it's great 696 00:45:06,030 --> 00:45:11,820 if that's already sort of accessible to me, and then multiple clients of 697 00:45:11,820 --> 00:45:13,500 theirs have the same type of problems. 698 00:45:13,710 --> 00:45:16,110 So if AI be doing those tasks. 699 00:45:18,120 --> 00:45:20,580 You know, they might as well deliver them via AI and via 700 00:45:20,580 --> 00:45:21,839 software to all those clients. 701 00:45:22,410 --> 00:45:23,520 Yeah, that makes sense. 702 00:45:23,700 --> 00:45:24,930 We're actually excited. 703 00:45:25,020 --> 00:45:29,310 Um, I mean, Harvey's announced a, a similar like client facing, you 704 00:45:29,310 --> 00:45:31,950 know, again, we get asked a lot, Hey, are you worried about these 705 00:45:32,009 --> 00:45:37,290 well-funded players coming in and talking about legal collaboration? 706 00:45:37,710 --> 00:45:42,150 And we're ex actually excited about it because it has been a very historically. 707 00:45:43,080 --> 00:45:48,810 Neglected space, like, I mean, the dominant player there is hiq that was 708 00:45:48,810 --> 00:45:50,850 bought by Thomson Reuters in 2019. 709 00:45:51,120 --> 00:45:55,710 The product hasn't changed a whole lot in, you know, the last seven years and 710 00:45:55,740 --> 00:45:59,640 there's only been one player in town and they own half the market share. 711 00:46:00,150 --> 00:46:05,070 So I, you know, I like to take an abundance mindset to thinking 712 00:46:05,070 --> 00:46:08,400 through this and I think it, it will create opportunity. 713 00:46:08,400 --> 00:46:11,880 Maybe at some point info dash will have integrations with. 714 00:46:13,155 --> 00:46:15,525 Lara and other AI platforms. 715 00:46:16,095 --> 00:46:19,485 Um, so yeah, I'm an, I'm an optimist. 716 00:46:19,485 --> 00:46:20,115 I think that. 717 00:46:21,430 --> 00:46:24,735 I think it's been a, a, a place that's been historically neglected. 718 00:46:25,125 --> 00:46:26,415 I think that law firms. 719 00:46:26,715 --> 00:46:31,005 And then, uh, this is the last question I'll ask you 'cause um, I know we're 720 00:46:31,005 --> 00:46:35,055 running over, but I think that law firm clients are gonna want different 721 00:46:35,085 --> 00:46:36,735 information from their law firms. 722 00:46:37,125 --> 00:46:43,300 So when you can turn around documents in hours or sometimes minutes, like 723 00:46:43,300 --> 00:46:48,825 you're gonna start measuring your law firm partners based on turn time on. 724 00:46:49,455 --> 00:46:55,425 Number of defects, um, you know, volume there, it, whereas if it takes a week 725 00:46:55,425 --> 00:47:00,345 like that, that's too slow to have a dashboard around how quickly legal 726 00:47:00,345 --> 00:47:02,085 processes are being turned around. 727 00:47:02,295 --> 00:47:06,495 So we think that there, there will be an appetite for. 728 00:47:07,575 --> 00:47:11,985 Inside legal teams to see metrics around these sorts of things. 729 00:47:11,985 --> 00:47:13,695 I don't do, do you share that view? 730 00:47:13,905 --> 00:47:13,965 Yeah. 731 00:47:13,965 --> 00:47:17,535 Well, I think like there's different categories of work and uh, like probably 732 00:47:17,535 --> 00:47:19,215 some, some stuff fit within that. 733 00:47:19,215 --> 00:47:21,705 And then there's gonna be very large complex matters. 734 00:47:21,705 --> 00:47:25,335 We're gonna be working for a longer time and still it's gonna be very 735 00:47:25,335 --> 00:47:29,055 useful to collaborate, um, with the enterprise or the customer that 736 00:47:29,055 --> 00:47:30,645 you're working with and the firm. 737 00:47:31,005 --> 00:47:33,495 And I think. 738 00:47:35,190 --> 00:47:39,150 It's the most exciting time in history to be a lawyer because, and a legal 739 00:47:39,150 --> 00:47:45,029 professional because no lo you know, never have there been so many talented 740 00:47:45,029 --> 00:47:48,240 people around the world working on building delightful software. 741 00:47:48,540 --> 00:47:53,670 And the pace of that development is staggering and a lot of fun. 742 00:47:53,940 --> 00:47:59,549 And just as you said, I'm also very, um, optimistic about the future world 743 00:47:59,549 --> 00:48:02,370 of abundance and how everybody's going to collaborate within its. 744 00:48:02,790 --> 00:48:03,210 Yeah. 745 00:48:03,629 --> 00:48:03,899 Yeah. 746 00:48:03,899 --> 00:48:11,430 It's easy to, you know, there have been lots of revolutions and transformations 747 00:48:12,029 --> 00:48:17,009 throughout human history, and going into them is always scary because 748 00:48:17,009 --> 00:48:19,710 change is hard, but I, I'm with you. 749 00:48:20,535 --> 00:48:23,895 People ask all the time, are you worried about somebody vibe, coding info dash? 750 00:48:24,315 --> 00:48:30,345 Which makes me chuckle because, you know, if they only knew how many, like how much 751 00:48:30,345 --> 00:48:34,245 learning it took, it's not the writing the code piece, that's the easy part. 752 00:48:34,635 --> 00:48:37,785 It's the learning, oh, you know what, that's the wrong API because 753 00:48:38,055 --> 00:48:41,985 that one gives you something that's different than what you actually need. 754 00:48:41,985 --> 00:48:43,605 I need this API over here. 755 00:48:44,115 --> 00:48:47,955 Um, so yeah, I, I, I think we're. 756 00:48:48,600 --> 00:48:54,840 We're headed for an era of abundance and I take much, I lean much more on the 757 00:48:54,840 --> 00:49:00,120 utopian, um, side than the dystopian, but I do think it's gonna be a bumpy ride. 758 00:49:00,509 --> 00:49:02,190 It's not gonna be smooth sailing. 759 00:49:02,430 --> 00:49:02,610 Yeah. 760 00:49:03,120 --> 00:49:07,380 Change is hard and um, I think there will be some bumps along 761 00:49:07,380 --> 00:49:08,490 the way, but we'll figure him out. 762 00:49:10,530 --> 00:49:10,945 I think so too. 763 00:49:11,835 --> 00:49:12,404 Awesome. 764 00:49:12,645 --> 00:49:14,265 Well, this has been a great conversation. 765 00:49:14,265 --> 00:49:17,595 I really appreciate you taking the time and sticking with me. 766 00:49:17,595 --> 00:49:20,265 We've been trying to get this on the books for at least a year. 767 00:49:20,564 --> 00:49:21,794 We finally got it done. 768 00:49:21,855 --> 00:49:25,814 Um, so I, I really appreciate you, uh, spending some time today. 769 00:49:27,075 --> 00:49:27,379 Thank you so much. 770 00:49:27,379 --> 00:49:28,665 Uh, that was fantastic. 771 00:49:28,875 --> 00:49:29,660 All right, well chat soon. 772 00:49:31,770 --> 00:49:32,160 Take care. 773 00:49:32,279 --> 00:49:34,560 Thanks for listening to Legal Innovation Spotlight. 774 00:49:35,100 --> 00:49:38,580 If you found value in this chat, hit the subscribe button to be notified 775 00:49:38,580 --> 00:49:40,080 when we release new episodes. 776 00:49:40,589 --> 00:49:43,259 We'd also really appreciate it if you could take a moment to rate 777 00:49:43,259 --> 00:49:45,899 us and leave us a review wherever you're listening right now. 778 00:49:46,470 --> 00:49:49,200 Your feedback helps us provide you with top-notch content. 00:00:01,680 How are you this afternoon? 2 00:00:02,340 --> 00:00:03,150 I'm amazing, Ted. 3 00:00:03,150 --> 00:00:03,870 It's so great to see you. 4 00:00:04,019 --> 00:00:05,190 Yeah, you as well. 5 00:00:05,490 --> 00:00:07,020 We've been trying to do this for. 6 00:00:07,860 --> 00:00:08,820 A while. 7 00:00:09,360 --> 00:00:13,260 Yeah, the schedule has been hectic for all the right reasons, but 8 00:00:13,260 --> 00:00:15,330 I'm really excited that we finally got the chance to sit down. 9 00:00:15,480 --> 00:00:21,119 Yeah, so am I. So am I. It's um, it's a conversation that's overdue, but 10 00:00:21,570 --> 00:00:22,680 I'm really looking forward to it. 11 00:00:22,680 --> 00:00:24,390 We have a really good agenda. 12 00:00:25,080 --> 00:00:27,210 Talk through lots, lots to talk through. 13 00:00:27,790 --> 00:00:31,270 Um, before we jump in, why don't you just do a quick introduction. 14 00:00:31,270 --> 00:00:34,120 I'm sure most people know who you are, but, um, for those that 15 00:00:34,120 --> 00:00:37,270 don't, just a little background on who you are, what you do Yeah. 16 00:00:37,270 --> 00:00:37,930 And where you do it. 17 00:00:38,590 --> 00:00:39,160 Of course. 18 00:00:39,280 --> 00:00:41,755 So, I'm Max, I'm the CO and co-founder of Lara. 19 00:00:42,010 --> 00:00:44,290 Lara is an AI platform for legal work. 20 00:00:44,290 --> 00:00:47,769 And the way that we've really started to think about it is we're building 21 00:00:47,769 --> 00:00:52,210 the place where both humans and agents will be completing legal work together. 22 00:00:52,215 --> 00:00:52,475 And so. 23 00:00:53,385 --> 00:00:57,435 We founded in 2023, just completed our Series D fundraise, where 24 00:00:57,435 --> 00:01:00,105 we raised over $550 million. 25 00:01:00,375 --> 00:01:04,335 Uh, we're about 400 people globally across eight offices, and we're 26 00:01:04,335 --> 00:01:05,325 having the time of our lives. 27 00:01:06,165 --> 00:01:07,065 It's incredible. 28 00:01:07,065 --> 00:01:08,055 It's incredible growth. 29 00:01:08,055 --> 00:01:12,405 Like you and I were on a podcast, I don't know if you remember this like I do. 30 00:01:12,435 --> 00:01:12,825 Yeah. 31 00:01:12,825 --> 00:01:13,245 Like this. 32 00:01:13,245 --> 00:01:16,020 That was, I, you, I, I think you guys were just starting up totally a year ago. 33 00:01:16,020 --> 00:01:16,300 A year and a half. 34 00:01:16,615 --> 00:01:19,255 I think it was more like two, two and a half years ago. 35 00:01:19,255 --> 00:01:23,664 I don't even know if I had my podcast then it was, it was early days. 36 00:01:23,905 --> 00:01:25,015 Well, you had a good setup at least. 37 00:01:25,164 --> 00:01:29,485 Yeah, and it's been it, man, it's been a lot of fun to watch, uh, 38 00:01:29,574 --> 00:01:33,774 you know, watch the whole space grow and watch you guys grow and. 39 00:01:34,330 --> 00:01:38,230 I've seen you've hired a lot of really good people on your team. 40 00:01:38,380 --> 00:01:41,080 Um, we had Kyle Poe one back in December. 41 00:01:41,080 --> 00:01:44,230 Yeah, I, I, I like to think that my job at this point, Ted, is to hire 42 00:01:44,230 --> 00:01:45,760 people much smarter than myself. 43 00:01:45,790 --> 00:01:49,750 And, uh, I've certainly had some pinch me moments where we do a 44 00:01:49,750 --> 00:01:52,630 company-wide town hall and you sort of look out over the group and you 45 00:01:52,630 --> 00:01:57,850 feel like, how on earth did I manage to, um, sort of lure everybody here. 46 00:01:58,180 --> 00:01:59,200 But no, we're grateful. 47 00:01:59,200 --> 00:02:00,010 We're super excited. 48 00:02:00,990 --> 00:02:04,470 Really, I think that's a big part of what, what keeps me going, uh, getting 49 00:02:04,470 --> 00:02:05,880 to work with so many inspiring people. 50 00:02:06,390 --> 00:02:06,780 Yeah. 51 00:02:06,840 --> 00:02:07,230 You know? 52 00:02:07,260 --> 00:02:11,880 Uh, have you ever read the book or heard of the book Good To Great by Jim Collins? 53 00:02:12,600 --> 00:02:13,380 No, but maybe I should. 54 00:02:13,380 --> 00:02:13,770 Yeah. 55 00:02:13,770 --> 00:02:15,540 It's a good, it's one worth reading. 56 00:02:15,570 --> 00:02:21,120 He talks about, uh, he, he goes back and does a study of, I don't know, 57 00:02:21,420 --> 00:02:25,080 hundreds of Fortune 500 companies over really long periods of time. 58 00:02:25,769 --> 00:02:31,200 He finds patterns and anti-patterns for success over the long haul. 59 00:02:31,559 --> 00:02:37,769 And one of the anti-patterns he calls the model of a genius in a thousand 60 00:02:37,769 --> 00:02:44,700 helpers, where, you know, one person, you know, may usually the CEO is, you 61 00:02:44,700 --> 00:02:49,674 know, kind of the, the maestro and the smartest guy in the room and it. 62 00:02:50,475 --> 00:02:56,055 It creates a lot of friction towards scale and um, yeah, it's a great book. 63 00:02:56,205 --> 00:02:59,085 Definitely worth I'll, I'll, I'll pick it up and, uh, that's 64 00:02:59,085 --> 00:03:00,915 certainly not Lara, right? 65 00:03:01,065 --> 00:03:02,655 Yeah, it's not info dash either. 66 00:03:04,425 --> 00:03:09,105 Um, well, why don't we start off with just talking about like the landscape, 67 00:03:09,165 --> 00:03:11,745 which is constantly changing and Yeah. 68 00:03:12,285 --> 00:03:15,975 You know, we had, it's been such a rollercoaster with like the 69 00:03:15,975 --> 00:03:18,015 Claude Legal plugin this year and. 70 00:03:18,720 --> 00:03:23,040 Um, you know, the whole, all this talk of rappers, which. 71 00:03:23,520 --> 00:03:27,540 I'm surprised that nobody's accused us of being a rapper on top of 72 00:03:27,540 --> 00:03:31,320 SharePoint, which, which we are essentially, like, we're built on 73 00:03:31,320 --> 00:03:34,290 top of SharePoint online and Azure. 74 00:03:34,890 --> 00:03:38,490 Um, nobody's called us a rapper yet, so I guess that's a good thing. 75 00:03:39,150 --> 00:03:41,550 Well, I think we're all, we're, we're all wrapping something, right? 76 00:03:42,060 --> 00:03:42,420 Yeah. 77 00:03:42,420 --> 00:03:45,840 You know, I mean, I've heard, I've heard the metaphor. 78 00:03:45,870 --> 00:03:48,510 Well, all apps are just rappers on a database. 79 00:03:49,049 --> 00:03:54,780 And I guess that's true to a certain extent, but, um, how do you, how do you 80 00:03:54,780 --> 00:04:00,269 see the landscape just broadly and where Lara fits in, like, you know, how, what 81 00:04:00,269 --> 00:04:03,060 makes Lara different from its competitors? 82 00:04:03,060 --> 00:04:04,019 Why don't we start with that? 83 00:04:04,829 --> 00:04:05,160 Sure. 84 00:04:05,579 --> 00:04:08,040 Well, look, it's, it's, you know. 85 00:04:08,565 --> 00:04:12,975 End of quarter one of 2026 and the models are where they are. 86 00:04:13,065 --> 00:04:16,965 I think they made a real sort of step change in their capability 87 00:04:16,965 --> 00:04:20,475 and intelligence over Christmas with the Opus 4.5 and 4.6. 88 00:04:21,195 --> 00:04:26,355 And what Lara and where we live in the stack is really sort of the um, central 89 00:04:26,355 --> 00:04:30,975 workplace where a lawyer wakes up in the morning, they open up legum, they 90 00:04:30,975 --> 00:04:33,015 might fire off some work to our agent. 91 00:04:33,525 --> 00:04:35,655 Our agents will start to perform those things. 92 00:04:35,835 --> 00:04:38,680 They will then check in on it, and then we allow the collaboration 93 00:04:38,750 --> 00:04:40,080 both internally with the team. 94 00:04:40,680 --> 00:04:46,080 But also externally with other firms or with their clients and Inc. Increasingly, 95 00:04:46,080 --> 00:04:50,790 we're also seeing large legal departments from um, sort of insurance companies, 96 00:04:50,790 --> 00:04:54,630 banks, large tech companies starting to adopt these tools internally. 97 00:04:55,080 --> 00:05:00,120 So I think there was a lot of initial buzz in 20 24, 20 25 with, 98 00:05:00,210 --> 00:05:01,800 you know, what can this tech do? 99 00:05:02,910 --> 00:05:04,200 Sort to the point. 100 00:05:05,535 --> 00:05:10,815 Scale of adoption is moving fast, and every new tool and new capability 101 00:05:10,815 --> 00:05:15,225 that we add, we can very nicely sort of add to the arsenal of 102 00:05:15,225 --> 00:05:16,575 what our agent can go out and do. 103 00:05:17,655 --> 00:05:22,245 Yeah, it, it really does seem like the world changed in December in terms 104 00:05:22,245 --> 00:05:28,335 of, you know, I think the, to your point, Opus 4.5 and then Opus 4.6, it 105 00:05:28,335 --> 00:05:32,805 really feel, it felt like there was a seismic shift in terms of capability. 106 00:05:34,305 --> 00:05:35,505 Foundation models. 107 00:05:35,895 --> 00:05:40,635 How has that, how has that affected your business? 108 00:05:40,725 --> 00:05:43,515 Um, obviously you're built on top of these models. 109 00:05:43,965 --> 00:05:44,085 Yeah. 110 00:05:44,235 --> 00:05:48,825 But is there a, you know, a mindset of you gotta stay one step ahead 111 00:05:48,825 --> 00:05:50,565 of the sheriff, so to speak. 112 00:05:50,565 --> 00:05:55,575 In terms of building on, on top of that, do they ever overlap in terms of 113 00:05:55,575 --> 00:05:58,395 capabilities that you have in existence? 114 00:05:58,395 --> 00:05:59,050 How do you think about that? 115 00:06:00,135 --> 00:06:03,195 Well look, when we founded LaGuardia in 2023, there was really 116 00:06:03,195 --> 00:06:04,845 just one model to work off of. 117 00:06:05,175 --> 00:06:09,885 It was GPT-3 0.5, uh, which was being hosted on Azure, like you mentioned. 118 00:06:09,885 --> 00:06:14,085 And, um, you know, from that point we've seen models sort of come 119 00:06:14,085 --> 00:06:17,865 and go and every quarter we get new capabilities to work with. 120 00:06:18,405 --> 00:06:23,175 I think one job of, of our team here at Lara is how do we bring the latest and 121 00:06:23,175 --> 00:06:27,435 greatest and how do we sort of squeeze the most juice out of the latest models 122 00:06:27,705 --> 00:06:30,195 for our use cases and for our customers? 123 00:06:30,795 --> 00:06:33,225 Because as you said, Ted, it's moving really fast. 124 00:06:33,315 --> 00:06:37,095 These are large organizations that we work with, large law firms who are 125 00:06:37,095 --> 00:06:41,655 adopting Lara Enterprise wide, and that also means that we need to bring those 126 00:06:41,655 --> 00:06:43,815 capabilities for different practice areas. 127 00:06:44,205 --> 00:06:47,085 So the way that you work with Legian litigation is different from m and 128 00:06:47,085 --> 00:06:48,614 a and it's different from in-house. 129 00:06:48,974 --> 00:06:54,525 And I think the models themselves are a sort of equalizer. 130 00:06:54,825 --> 00:06:56,234 Everybody has access to them, right? 131 00:06:56,265 --> 00:06:59,895 It's sort of the, like the new electricity, but you still need 132 00:06:59,895 --> 00:07:04,664 to put that electricity in a, in a great car, um, or, you know, sort 133 00:07:04,664 --> 00:07:06,494 of get, get the most outta them. 134 00:07:06,854 --> 00:07:09,164 And the way that I like to think about Lago is like a Formula 135 00:07:09,164 --> 00:07:12,015 One car, and then we can hot swap the engine anytime we want. 136 00:07:12,795 --> 00:07:16,905 So the capabilities that the models themselves solve, I 137 00:07:16,905 --> 00:07:17,955 think is ever increasing. 138 00:07:18,405 --> 00:07:20,565 That just allows us to build more on top of them. 139 00:07:21,615 --> 00:07:22,275 Interesting. 140 00:07:22,664 --> 00:07:28,755 And then the, you know, SaaS apocalypse as they call it, um, the Claude 141 00:07:28,755 --> 00:07:33,885 legal plugin, which I have a lot of thoughts on that I've, some of which 142 00:07:33,885 --> 00:07:35,865 I've shared on previous podcasts. 143 00:07:35,865 --> 00:07:41,655 I did a podcast with Ilta recently where we, um, myself, Kat Casey and Dan Zebo. 144 00:07:42,405 --> 00:07:47,805 Um, did a deep dive into what the meaning of it really is. 145 00:07:47,805 --> 00:07:53,895 And, you know, my theory on why there was such a swift market 146 00:07:53,895 --> 00:07:57,735 reaction to really, there were no new capabilities as part of the legal 147 00:07:57,735 --> 00:07:59,775 plugin, like absolutely nothing new. 148 00:08:00,105 --> 00:08:01,635 The plugin architecture existed. 149 00:08:02,145 --> 00:08:07,305 It's some really simple markdown with, you know, six skills that are extremely basic. 150 00:08:07,305 --> 00:08:08,655 There were no new capabilities. 151 00:08:09,015 --> 00:08:14,445 So why did it have such a. A huge impact on the markets. 152 00:08:15,045 --> 00:08:20,145 My theory on that, and I'd be curious to get yours, my theory on that is, 153 00:08:21,525 --> 00:08:28,845 um, the plugin was a release valve for a lot of anxiety in the marketplace 154 00:08:28,905 --> 00:08:37,034 over a lot of capital investment, high valuations, uncertainty about how. 155 00:08:37,455 --> 00:08:42,255 You know how AI is going to impact certain industries, and you saw 156 00:08:42,255 --> 00:08:47,835 companies impacted that shouldn't have been like, you know, TR and Lexus with 157 00:08:47,835 --> 00:08:50,295 all of the proprietary data they have. 158 00:08:50,715 --> 00:08:53,415 LegalZoom got hit hard and that made more sense to me. 159 00:08:53,685 --> 00:08:57,315 I actually see more impact to a LegalZoom business model 160 00:08:57,315 --> 00:08:59,505 than I would to TR or Lexus. 161 00:09:00,135 --> 00:09:02,865 Um, but I, it, it felt like a release valve. 162 00:09:03,355 --> 00:09:07,855 And an excuse for this anxiety to manifest itself. 163 00:09:08,275 --> 00:09:08,545 I don't know. 164 00:09:08,545 --> 00:09:12,235 What is your take on the legal plugin and its impact on the market? 165 00:09:14,155 --> 00:09:22,525 Well, I think you have traditional sauce companies that are more or less 166 00:09:23,155 --> 00:09:29,905 well equipped to execute on what an AI native stack and future will look like. 167 00:09:31,105 --> 00:09:35,280 I think that's independent or sort of, you know, set aside from, 168 00:09:35,730 --> 00:09:39,390 you know, whether, uh, Claude can work with a legal skill or not. 169 00:09:39,390 --> 00:09:39,690 Right. 170 00:09:40,200 --> 00:09:44,460 I think you're somewhat right in that, you know, maybe it was a, a marketable moment 171 00:09:44,460 --> 00:09:48,630 that became an opportunity to, or a reason to sort of, you know, pivot a certain way. 172 00:09:49,650 --> 00:09:54,780 Now I think generally the source that has been built up, like the cost 173 00:09:54,780 --> 00:09:57,240 of writing software is going down. 174 00:09:59,579 --> 00:10:03,270 Opportunity space and the amount of greenfield problems you can go and 175 00:10:03,270 --> 00:10:07,920 solve have actually significantly increased with the model capability. 176 00:10:08,579 --> 00:10:12,209 So we had this really funny moment because we were out racing our series D 177 00:10:12,719 --> 00:10:18,180 at the time of this release, and you had the public, you know, stock sell off, 178 00:10:18,780 --> 00:10:23,880 and we experienced astronomical demand in our own private fundraising round. 179 00:10:24,329 --> 00:10:24,719 And so. 180 00:10:25,770 --> 00:10:31,530 Although there's may, maybe a, a lack of, of trust and excitement around what the 181 00:10:31,589 --> 00:10:35,520 public sales companies can do in terms of executing on this AI native future, 182 00:10:36,120 --> 00:10:40,380 there was a real excitement about, uh, ours and a few, you know, other companies 183 00:10:40,380 --> 00:10:41,699 ability to go and execute on that. 184 00:10:41,699 --> 00:10:45,839 So now I think there's trade-offs, um, and 185 00:10:47,880 --> 00:10:49,020 I am. 186 00:10:50,640 --> 00:10:57,270 Again, sort of very happy that the models keep developing because data is also 187 00:10:57,270 --> 00:11:01,320 playing into the way that our strategy is, which is let's sort of build, you 188 00:11:01,320 --> 00:11:06,420 know, every functionality in Lago like a boat, and when the tide rises, when the 189 00:11:06,420 --> 00:11:11,430 underlying capabilities improve, every single feature that we have gets better. 190 00:11:12,390 --> 00:11:17,280 And I also tried to tell the team, Ted, that we need to build with a very low ego. 191 00:11:18,090 --> 00:11:21,630 You might spend a lot of time building something, you know, nine months ago 192 00:11:21,840 --> 00:11:26,820 that made a lot of sense back then, but it won't make so much sense now that 193 00:11:26,820 --> 00:11:31,170 we're in a world where agents will be executing and doing work on our behalf. 194 00:11:31,680 --> 00:11:31,800 Right. 195 00:11:31,800 --> 00:11:36,420 I think we've moved from the sort of co-pilot, you know, let's iterate 196 00:11:36,420 --> 00:11:40,170 very tightly to, I'm actually gonna let the agent go and work 197 00:11:40,170 --> 00:11:43,080 for a bit and then review the output that it's coming back with. 198 00:11:44,339 --> 00:11:47,699 Yeah, and you know, that was another theory that was put forward in 199 00:11:47,699 --> 00:11:54,780 terms of why the selloff took place, which is with how good agents have 200 00:11:54,780 --> 00:11:59,670 gotten, even automating a ui, right? 201 00:11:59,760 --> 00:12:04,110 That it's going to put pressure on the per seat licensing model. 202 00:12:04,290 --> 00:12:08,040 If you can license an agent to do what three people can do, you 203 00:12:08,040 --> 00:12:10,079 now need one third, the licenses. 204 00:12:10,920 --> 00:12:15,015 And I, you know, I. Like, okay. 205 00:12:15,045 --> 00:12:17,954 But we can adapt to that. 206 00:12:17,954 --> 00:12:22,545 We can, but that's not the only licensing model that's available in the universe. 207 00:12:22,545 --> 00:12:22,785 Right. 208 00:12:22,785 --> 00:12:25,725 We can, that you can get, you can get around that. 209 00:12:26,175 --> 00:12:34,755 Um, how do you see the, you know, agentic phenomenon impacting per user licensing? 210 00:12:34,755 --> 00:12:37,814 'cause I would imagine you're on a per seat license model as well. 211 00:12:37,814 --> 00:12:38,079 Is that accurate? 212 00:12:39,285 --> 00:12:42,675 Yeah, we're on a per seat and, and sometimes consumption as well. 213 00:12:42,735 --> 00:12:48,675 And I think, look, you, you've had, uh, tools for engineers that have been 214 00:12:48,675 --> 00:12:54,165 out for, for a time, if you look at Cursor or sort of the first, um, really 215 00:12:54,165 --> 00:12:58,305 adjunct tools that came out and they've always been consumption based pricing. 216 00:12:58,305 --> 00:13:02,835 And so I think that as a paradigm makes sense for sure. 217 00:13:03,495 --> 00:13:07,615 Um, and then I think there's a question of how do we, as a market, you know. 218 00:13:09,314 --> 00:13:13,995 So that we can control that in an effective way, right? 219 00:13:13,995 --> 00:13:18,735 Like, um, if you're leveraging a lot of AI in an e-discovery process, 220 00:13:19,155 --> 00:13:22,520 you're already used to paying something on a consumption based model. 221 00:13:23,204 --> 00:13:25,935 'cause you could run, you know, hundreds of millions of documents. 222 00:13:26,564 --> 00:13:30,135 And so I think there's a, there is an, an understanding of that. 223 00:13:30,255 --> 00:13:34,574 And as we move more and more into these agent, uh, sort of task 224 00:13:34,574 --> 00:13:38,415 completions, as you explained Ted, you are using more and more tokens. 225 00:13:39,045 --> 00:13:43,485 And so in order for that margin profile and for that, uh, work to actually 226 00:13:43,485 --> 00:13:47,655 make sense, I think there has to be some level of consumption, uh, pricing 227 00:13:47,714 --> 00:13:52,814 associated with it is, is outcome. 228 00:13:53,670 --> 00:14:02,189 Or deliverable based pricing in, in the near term, or is that still further away? 229 00:14:02,939 --> 00:14:07,230 I think that's possible, but I, it's, it's, it's harder because then 230 00:14:07,230 --> 00:14:10,050 you need, you need to build out the taxonomy of the task that you can 231 00:14:10,050 --> 00:14:12,569 solve, right, to a very granular level. 232 00:14:12,660 --> 00:14:17,490 And the different types of tasks that you can solve with leggo, at least 233 00:14:17,490 --> 00:14:20,370 across the different practice areas is really broad and really varied. 234 00:14:22,500 --> 00:14:27,210 We're also seeing firms moving to that from the billable hour, right? 235 00:14:27,210 --> 00:14:30,330 In certain practice areas and in certain geographies. 236 00:14:30,360 --> 00:14:34,920 And so I think all of these developments, the waterfall will sort of happen 237 00:14:35,640 --> 00:14:37,320 slowly, but sort of altogether. 238 00:14:38,730 --> 00:14:44,189 And then I, um, I'm sure you heard of about the, the post on 239 00:14:44,189 --> 00:14:48,870 XI think it was Zach Shapiro who wrote about the Claude Native Firm. 240 00:14:49,560 --> 00:14:55,860 Um, and there's been a lot of back and forth on LinkedIn about, you know, how, 241 00:14:55,920 --> 00:15:03,000 how realistic, um, some of what was put forward actually is in terms of how do 242 00:15:03,000 --> 00:15:05,970 you, how, how do you explain to someone. 243 00:15:06,780 --> 00:15:12,989 What the difference is between leveraging just, you know, 244 00:15:13,140 --> 00:15:17,459 clawed and a bunch of prompts and skills versus leveraging a leg. 245 00:15:18,089 --> 00:15:21,030 How do you differentiate it to people? 246 00:15:21,060 --> 00:15:23,489 How, how, what those different paths look like? 247 00:15:25,074 --> 00:15:29,369 Um, I, I've, I've spent enough time with lawyers to learn to say that it depends, 248 00:15:29,369 --> 00:15:31,890 but it really depends on the problem that you're trying to solve, right? 249 00:15:32,520 --> 00:15:33,750 Let's say you are. 250 00:15:36,750 --> 00:15:38,220 Contracting in an enterprise. 251 00:15:39,390 --> 00:15:43,950 The way that you actually set that up is you want the entire, you know, 252 00:15:43,950 --> 00:15:48,390 company to be able to leverage the same set of playbooks and, and uh, 253 00:15:48,450 --> 00:15:50,010 work, produce by the legal team. 254 00:15:50,760 --> 00:15:53,700 And then you want an agent to go and review all those documents, 255 00:15:53,790 --> 00:15:57,150 create markup, and sort of be ready to send back to the other party. 256 00:15:57,600 --> 00:15:59,220 And you need an example language. 257 00:15:59,220 --> 00:16:01,710 You want different fallbacks for different types of documents. 258 00:16:01,980 --> 00:16:04,510 There's a lot of routing in takeaway that you actually work with that. 259 00:16:06,390 --> 00:16:10,590 Or the model as the engine, again, I think makes a lot of sense. 260 00:16:10,590 --> 00:16:14,310 It's the electricity, but then it's the house or the car that you actually 261 00:16:14,310 --> 00:16:18,030 build around that that becomes valuable for an enterprise to deploy. 262 00:16:18,900 --> 00:16:22,800 And if you look on an m and a transaction, the way that you work with the data 263 00:16:22,800 --> 00:16:25,590 room and the way that you sort of walk through that, the way that you then 264 00:16:25,590 --> 00:16:30,360 take the IP part and the employment part, and the finance or tax part, and 265 00:16:30,360 --> 00:16:33,960 then you bundle all of that together into a single due diligence report. 266 00:16:35,040 --> 00:16:39,300 All the work around where the model is actually working is the stuff that we do. 267 00:16:39,900 --> 00:16:44,400 And then we work really hard to, as, as I said, like get the maximum 268 00:16:44,400 --> 00:16:49,439 juice out of the model so that you can trust that the output that you 269 00:16:49,439 --> 00:16:51,990 get back is based in the context. 270 00:16:52,140 --> 00:16:55,980 With citations, you can market, you can red flag it, you could green 271 00:16:55,980 --> 00:16:59,250 light it, and then everybody can collaborate around that at the same time. 272 00:17:00,030 --> 00:17:03,660 And I think the more time goes on, the bigger and bigger difference. 273 00:17:04,500 --> 00:17:08,940 La Agora has from the foundational models, to be honest with you, how, 274 00:17:09,180 --> 00:17:15,480 how do you think law firms are going to differentiate themselves in this 275 00:17:15,480 --> 00:17:17,640 post, in the post transformation era? 276 00:17:17,970 --> 00:17:25,140 Like historically differentiation, like I actually spun up Claude Cowork and as 277 00:17:25,140 --> 00:17:32,520 Input gave it the AM law 100 list and had it go and list out how firms frame. 278 00:17:33,645 --> 00:17:35,745 Their differentiation, narrative. 279 00:17:36,255 --> 00:17:42,495 And, um, eight out of 100 even mentioned technology, right? 280 00:17:42,615 --> 00:17:43,755 Pretty low percentage. 281 00:17:43,755 --> 00:17:48,195 Historically, how law firms have differentiated themselves is, you know, 282 00:17:48,225 --> 00:17:56,145 their people, their brand, their their wins, their, you know, um, their history. 283 00:17:56,805 --> 00:18:00,405 And as we get into a, taking a tech enabled world. 284 00:18:01,770 --> 00:18:06,150 Technology is going to be part of that differentiation equation, but buying off 285 00:18:06,150 --> 00:18:11,880 the shelf tools isn't differentiation, isn't differentiating by itself. 286 00:18:12,210 --> 00:18:15,900 So how do you think, you know, again, kind of post transformation, 287 00:18:15,960 --> 00:18:19,230 and I don't know when that is gonna be, but we're probably gonna be 288 00:18:19,230 --> 00:18:21,210 transforming for a very long time. 289 00:18:21,630 --> 00:18:25,980 But how are firms going to differentiate themselves from each other? 290 00:18:25,980 --> 00:18:27,300 If everybody's running Lara? 291 00:18:28,350 --> 00:18:30,899 How, how am I different than from the firm down the street? 292 00:18:31,889 --> 00:18:37,110 Well, ever great architect works in cud and ever great designer works 293 00:18:37,110 --> 00:18:42,120 in Figma and every great engineer works in cursor or cloud code. 294 00:18:42,270 --> 00:18:45,294 And so, um, you know, if I get a paint brush. 295 00:18:46,379 --> 00:18:48,179 I could do a pretty shitty self portrait. 296 00:18:48,570 --> 00:18:52,230 If I gave a paintbrush to Michelangelo, it would be an amazing piece of art. 297 00:18:52,830 --> 00:18:56,580 And so the tools is of course, one part of the equation. 298 00:18:56,790 --> 00:19:01,919 Are you equipping your people and your firm with the prerequisites 299 00:19:02,370 --> 00:19:06,300 to go out on the, you know, metaphorical battlefield and win? 300 00:19:06,420 --> 00:19:10,710 You know, you don't wanna show up with a plastic sword and you know, your 301 00:19:10,710 --> 00:19:15,930 opponent has a, has a, you know, metal on, uh, then you're not equipping the 302 00:19:15,930 --> 00:19:17,910 firm with the opportunity to go and win. 303 00:19:18,840 --> 00:19:23,430 And I think la, you know, equivalent, like that's part of, of this stack 304 00:19:23,430 --> 00:19:25,560 required to actually do real work. 305 00:19:26,220 --> 00:19:30,420 Now, the way that you leverage those tools, Ted, that can 306 00:19:30,420 --> 00:19:31,920 be really differentiated. 307 00:19:32,595 --> 00:19:39,045 I know Lara users who are standard deviations ahead on how AI savvy they're, 308 00:19:39,855 --> 00:19:44,955 and the amount of work that they can get done at the quality that they do it with. 309 00:19:44,955 --> 00:19:48,075 The help of systems like Lara is amazing. 310 00:19:48,375 --> 00:19:49,485 It's, it's incredible. 311 00:19:49,545 --> 00:19:52,155 It's like what's watching a magician work. 312 00:19:52,875 --> 00:19:57,225 But then I also know people who, uh, open up La Gora and they give it a 313 00:19:57,225 --> 00:19:59,446 prompt, which is, write me an SPA. 314 00:20:01,274 --> 00:20:05,835 They're unsatisfied with the output and they go, Hey, I can't do my job. 315 00:20:06,345 --> 00:20:07,425 I'm not gonna look at it. 316 00:20:08,325 --> 00:20:10,935 I think that's a very unsuccessful strategy. 317 00:20:11,504 --> 00:20:15,405 And so even if your firm has Lara, if you don't know how to use it, it's 318 00:20:15,405 --> 00:20:16,754 not gonna be very valuable for you. 319 00:20:17,290 --> 00:20:21,794 And so what I think will be a real differentiator is, one, 320 00:20:22,034 --> 00:20:23,205 how you set up the system. 321 00:20:23,835 --> 00:20:27,495 Two, how you enable your people to actually leverage the system. 322 00:20:28,095 --> 00:20:28,965 And then three. 323 00:20:29,820 --> 00:20:36,389 How, um, how much on the frontier you sort of keep pushing and how much do 324 00:20:36,389 --> 00:20:40,740 you think about your processes and how should we redo our m and a process to 325 00:20:40,740 --> 00:20:45,179 actually deliver a quicker turnaround for our client or a better quality work? 326 00:20:45,360 --> 00:20:48,240 Like that's where the differentiation is going to start to seep in 327 00:20:49,379 --> 00:20:51,659 with all of the other things that you also mentioned, right? 328 00:20:51,659 --> 00:20:53,820 The talent, the brand, the people. 329 00:20:54,090 --> 00:20:55,200 This is the whole package. 330 00:20:55,965 --> 00:20:56,445 Yeah. 331 00:20:56,504 --> 00:21:02,445 You know, I, I really think that the firms are going to have to figure out 332 00:21:02,445 --> 00:21:09,104 how to harness and leverage the knowledge and wisdom that led to winning outcomes 333 00:21:09,104 --> 00:21:12,165 for their clients along with ai, right? 334 00:21:12,195 --> 00:21:12,314 Yeah. 335 00:21:12,375 --> 00:21:16,004 That is a lot of times locked up in a very messy. 336 00:21:16,560 --> 00:21:22,320 DMS, um, you know, there are companies out there that are, you know, cro. 337 00:21:22,650 --> 00:21:26,280 Everybody goes, I can never find what I need in the document management system. 338 00:21:26,640 --> 00:21:30,600 And the reason is because there's a terrible signal to noise ratio. 339 00:21:31,080 --> 00:21:31,860 You know, you've got, you. 340 00:21:32,399 --> 00:21:38,760 17 versions of the same document with Finalco Finalco final 341 00:21:39,000 --> 00:21:40,740 on the end of the file name. 342 00:21:41,219 --> 00:21:43,865 It's very difficult to get in there and make sense of you. 343 00:21:43,865 --> 00:21:46,860 You, you know, you've got emails, you've got matters that, 344 00:21:47,010 --> 00:21:49,320 um, never actually transacted. 345 00:21:49,919 --> 00:21:51,665 Uh, but there's still all of that. 346 00:21:52,304 --> 00:21:57,254 Um, all of those documents in the DMS that if you just turn AI loose 347 00:21:57,585 --> 00:22:01,875 on a DMS, it's going to have the same struggles that a person is that's 348 00:22:01,875 --> 00:22:04,845 trying to find something, um, useful. 349 00:22:04,845 --> 00:22:07,335 That's actually, that actually created a winning outcome. 350 00:22:07,754 --> 00:22:13,754 So how do you think about how law firms are going to leverage that content 351 00:22:13,754 --> 00:22:19,064 and wisdom and the things that created winning outcomes in the AI future? 352 00:22:19,095 --> 00:22:20,235 What does that look like in your mind? 353 00:22:21,495 --> 00:22:25,005 Well, I think that's one part of the puzzle, but, but certainly not all of it. 354 00:22:25,395 --> 00:22:33,105 Um, and what I'd almost, you know, sort of rephrase it to is how do you provide an 355 00:22:33,105 --> 00:22:39,885 agent or your AI software with the context and the capability to fetch that context 356 00:22:40,395 --> 00:22:42,315 so that it can do real work for you? 357 00:22:43,065 --> 00:22:43,305 Right. 358 00:22:43,455 --> 00:22:44,235 And that context. 359 00:22:45,345 --> 00:22:50,205 In your dms, it can live in your email, it could live online, could live in a, 360 00:22:50,535 --> 00:22:53,145 you know, legal research, uh, repository. 361 00:22:54,135 --> 00:22:54,555 And 362 00:22:57,135 --> 00:23:04,575 as you say, navigating these systems by hand as a human is really inefficient and. 363 00:23:06,225 --> 00:23:10,965 Luckily for US, AI and software is a lot faster, and so you can actually let the 364 00:23:10,965 --> 00:23:17,625 ai, you know, query against the databases using either APIs or MZP, and that starts 365 00:23:17,625 --> 00:23:23,385 to get, you know, starts to create the ability for the agents to actually go 366 00:23:23,385 --> 00:23:24,705 and fetch the context that they need. 367 00:23:25,455 --> 00:23:29,024 But what you also then want is you want the orchestration engine and you 368 00:23:29,024 --> 00:23:30,524 want the layer on top that actually. 369 00:23:31,485 --> 00:23:35,354 It tells the agent, this is what you need to do in what order. 370 00:23:35,715 --> 00:23:38,715 This is your plan and this is your execution in order to actually 371 00:23:38,715 --> 00:23:40,215 produce the output that we want here. 372 00:23:40,574 --> 00:23:43,665 Whether that's fetching precedent or going out online to find something. 373 00:23:44,145 --> 00:23:47,385 So again, I think this is sort of, you know, the modern stack here 374 00:23:47,504 --> 00:23:48,705 is really starting to develop. 375 00:23:49,455 --> 00:23:50,054 How about. 376 00:23:50,385 --> 00:23:56,595 Like org structure and, and, um, how big a factor that is with success. 377 00:23:56,955 --> 00:24:00,765 So, for example, you know, I've been selling into the legal market for 20 378 00:24:00,765 --> 00:24:06,435 years and I, historically they have, from my perspective, under invested in 379 00:24:06,435 --> 00:24:11,445 technology and many times in knowledge management in general, knowledge 380 00:24:11,445 --> 00:24:12,975 management has become a catchall. 381 00:24:13,425 --> 00:24:13,995 It's like. 382 00:24:14,745 --> 00:24:19,965 You know, they do everything from actually true knowledge management or we call it 383 00:24:19,965 --> 00:24:25,544 Cam Classic, which is, you know, building and maintaining precedent libraries and, 384 00:24:25,875 --> 00:24:27,945 you know, base, base and model documents. 385 00:24:28,395 --> 00:24:34,004 Um, but they also have really been thrown into the fray to just become. 386 00:24:34,649 --> 00:24:36,450 Change management for, for it. 387 00:24:36,629 --> 00:24:40,050 We've got a new DMS coming and you know, you're gonna help us 388 00:24:40,050 --> 00:24:41,280 figure out how to organize it. 389 00:24:41,850 --> 00:24:44,850 Like what, what do you think about new roles in terms of, we're 390 00:24:44,850 --> 00:24:49,590 hearing all of this Palantir forward deployed engineer has become all 391 00:24:49,590 --> 00:24:51,810 the rage and, and legal engineer. 392 00:24:52,230 --> 00:24:56,220 Like from your perspective, the, the, the firms that have been most 393 00:24:56,220 --> 00:24:57,840 successful deploying your product. 394 00:24:58,290 --> 00:25:00,750 What sort of, what have they done differently from an 395 00:25:00,750 --> 00:25:01,950 organizational perspective? 396 00:25:02,429 --> 00:25:02,639 Yeah. 397 00:25:03,210 --> 00:25:08,610 Well, we were early, uh, to, to, to coin the term forward deployed legal engineer. 398 00:25:08,610 --> 00:25:12,480 And I think it's certainly a way that we've been able to work tightly with 399 00:25:12,480 --> 00:25:16,380 our clients from the very beginning to both co-create the product and 400 00:25:16,380 --> 00:25:18,480 make sure that we drive the adoption. 401 00:25:18,480 --> 00:25:20,790 That then leads to the ROI for these teams. 402 00:25:21,510 --> 00:25:21,570 Um. 403 00:25:23,250 --> 00:25:28,655 As you say, like we've seen every possible permutation of organizational design and, 404 00:25:28,660 --> 00:25:32,850 and team staffing from, from the firms and organizations that we work with. 405 00:25:32,850 --> 00:25:39,720 But generally, um, I think if you create the team whose role and 406 00:25:39,720 --> 00:25:44,760 responsibility is it is to drive innovation, you then need to give them 407 00:25:44,880 --> 00:25:48,150 enough political power and sort of, um. 408 00:25:51,975 --> 00:25:54,675 Opportunity to actually go and drive that change. 409 00:25:54,825 --> 00:25:59,715 Um, the most unsuccessful things we've seen is when you invest a ton 410 00:25:59,715 --> 00:26:03,284 of time in building up an amazing team, and then, you know, you 411 00:26:03,405 --> 00:26:07,215 just don't let them work directly with partners or drive the change. 412 00:26:07,875 --> 00:26:13,185 And, and then I think also most people underestimate the. 413 00:26:14,220 --> 00:26:20,370 Change that is required on an individual's sort of work level in order to 414 00:26:20,460 --> 00:26:23,040 properly get the maximum efficiency. 415 00:26:23,910 --> 00:26:27,840 And if you get a sort of one-on-one, you know, you start to learn how to 416 00:26:27,840 --> 00:26:29,850 work a bit with an assistant like. 417 00:26:30,185 --> 00:26:32,465 That's great, but that's sort of the baseline. 418 00:26:32,645 --> 00:26:37,445 Um, you need to take serious time out of your day to rethink how 419 00:26:37,445 --> 00:26:39,365 you're going to approach your work. 420 00:26:39,754 --> 00:26:45,004 Now, given the tools at your disposal and some change management teams and 421 00:26:45,004 --> 00:26:50,284 innovation teams and M teams, I think have been very effective in helping partners 422 00:26:50,284 --> 00:26:52,450 and practices navigate those changes. 423 00:26:53,165 --> 00:26:56,225 And they almost approach it as process engineers in a way. 424 00:26:56,375 --> 00:26:56,465 Mm-hmm. 425 00:26:56,705 --> 00:26:59,314 So they sort of map out like, okay, this is what we're doing manually. 426 00:26:59,685 --> 00:27:04,605 Now with this new tool, how could we approach it differently instead of going, 427 00:27:04,935 --> 00:27:08,745 okay, for all of these different, you know, here I might use a little bit of 428 00:27:08,745 --> 00:27:10,305 AI here, I might use a little bit of ai. 429 00:27:10,605 --> 00:27:12,165 They sort of approach it a bit more holistically. 430 00:27:12,764 --> 00:27:18,795 Yeah, and you know, historically legal work has done largely on a bespoke basis. 431 00:27:19,335 --> 00:27:20,475 You could go to the same. 432 00:27:21,659 --> 00:27:26,790 Law firm in the same practice and ask for a commercial lease within 433 00:27:26,790 --> 00:27:30,450 the real estate practice at the same law firm and get back four different. 434 00:27:31,230 --> 00:27:33,270 Results from four different attorneys, right? 435 00:27:33,540 --> 00:27:37,500 Because there's, do you know, everybody has their and, and that has been a result 436 00:27:37,530 --> 00:27:42,600 of lessons that these lawyers have learned along the way that they've incorporated 437 00:27:42,600 --> 00:27:47,580 these learnings into their own individual kind of bespoke work processes. 438 00:27:47,940 --> 00:27:51,570 And now we're entering into an era where we're becoming tech enabled. 439 00:27:51,990 --> 00:27:56,520 And this bespoke model is at odds with that if we're gonna scale. 440 00:27:57,360 --> 00:28:04,110 So how do you see, um, it seems like a herding of cats, uh, if you 441 00:28:04,110 --> 00:28:08,550 will, getting lawyers to standardize on their processes again, even 442 00:28:08,550 --> 00:28:10,260 within the same firms sometimes. 443 00:28:10,260 --> 00:28:12,064 So how have you seen that play out? 444 00:28:13,715 --> 00:28:14,804 Well, I 445 00:28:18,540 --> 00:28:21,689 think there's finally tools and opportunities that 446 00:28:21,689 --> 00:28:24,840 motivate doing that because. 447 00:28:25,830 --> 00:28:30,930 It's not just a linear improvement on a way you used to do something. 448 00:28:31,290 --> 00:28:37,020 This is a complete game changer in terms of how you're delivering your work, right? 449 00:28:37,260 --> 00:28:41,280 And it's like going from, you know, pen and paper to the computer. 450 00:28:41,280 --> 00:28:43,680 Like it's a big shift in terms of how you actually work. 451 00:28:43,740 --> 00:28:48,690 And I think the reality is that, um, you have real champions. 452 00:28:52,395 --> 00:28:57,315 Sort of really adopt early and show everybody else what's possible. 453 00:28:58,034 --> 00:29:01,665 And that needs to happen internally in order for the rest 454 00:29:01,665 --> 00:29:02,955 of the organization to follow. 455 00:29:03,554 --> 00:29:07,034 And so one thing that we often do when we're part of our sort of broader 456 00:29:07,034 --> 00:29:11,054 rollouts, but even in pilots, is that we really work with the firm 457 00:29:11,054 --> 00:29:13,185 to identify who this sort of, um. 458 00:29:15,045 --> 00:29:19,575 Fire carriers are gonna be because those are the people that we need to lift up 459 00:29:19,575 --> 00:29:23,835 internally and that we need to highlight because they're gonna be able to drag 460 00:29:23,835 --> 00:29:25,065 the rest of the organization with them. 461 00:29:25,695 --> 00:29:31,275 And I think increasingly Ted, there's also an expectation of their clients to 462 00:29:31,275 --> 00:29:35,925 start to leverage these tools to drive better value and better outcome for them. 463 00:29:36,405 --> 00:29:40,845 Because I actually don't no longer think that, that it's an excuse that, 464 00:29:40,850 --> 00:29:44,205 oh, oh, we haven't started looking at this, or, or something like that. 465 00:29:44,595 --> 00:29:51,885 Now, I think it's very hard to actually motivate not using AI for certain things. 466 00:29:52,425 --> 00:29:54,435 And even if you do things by hand. 467 00:29:55,139 --> 00:29:59,459 You should certainly do it with AI as well, because you might find 468 00:29:59,459 --> 00:30:00,570 things that you otherwise wouldn't. 469 00:30:00,990 --> 00:30:03,179 And so, or find strategies, right? 470 00:30:03,179 --> 00:30:07,530 Like on a, on a litigation case, you might even spend more time if you 471 00:30:07,530 --> 00:30:11,879 work with AI because you're gonna find out more creative strategies 472 00:30:12,149 --> 00:30:13,770 that you can potentially follow. 473 00:30:14,280 --> 00:30:14,669 And so. 474 00:30:16,110 --> 00:30:21,420 I think that there's gonna be sort of multiple classes of, of, of, of, um, 475 00:30:21,450 --> 00:30:23,100 sophistication that comes out here. 476 00:30:23,100 --> 00:30:28,470 But I'm certain that the, you know, really AI empowered and AI native 477 00:30:28,710 --> 00:30:30,000 lawyers will do extraordinarily well. 478 00:30:31,550 --> 00:30:32,840 Yeah, that makes a lot of sense. 479 00:30:32,900 --> 00:30:36,830 What do you think about the chief AI officer role at a law firm? 480 00:30:36,890 --> 00:30:39,380 You know, historically we've had chief innovation officers. 481 00:30:39,950 --> 00:30:46,730 Um, chief AI officer feels a little more narrow than, than just innovation. 482 00:30:46,850 --> 00:30:46,910 Yeah. 483 00:30:46,910 --> 00:30:51,560 I mean, look, I think AI is the biggest existential opportunity and 484 00:30:51,740 --> 00:30:56,150 maybe threats that's happened to the industry in a very, very long time. 485 00:30:56,480 --> 00:30:56,810 Um. 486 00:30:57,825 --> 00:31:04,004 I'm frankly surprised by how little resources many firms are allocating 487 00:31:04,215 --> 00:31:08,865 to this, um, which is understandable given the way that the, you know, 488 00:31:09,045 --> 00:31:10,754 profit sort of gets split every year. 489 00:31:11,475 --> 00:31:15,045 But then again, I think this is just something you have to get right, 490 00:31:15,615 --> 00:31:18,945 and not only do you have to get it right, if you're in a very competitive 491 00:31:18,945 --> 00:31:23,415 environment, you need to get it right more so than your competitors. 492 00:31:24,165 --> 00:31:28,695 And I think that requires real resources, real time, real energy and commitment 493 00:31:28,695 --> 00:31:32,895 from both your partner and the firm. 494 00:31:33,525 --> 00:31:34,755 And you have to work together on that. 495 00:31:35,355 --> 00:31:39,225 I think we've, you know, certainly felt that. 496 00:31:40,380 --> 00:31:43,830 In the biggest deployments that we've done and the tightest partnerships 497 00:31:43,830 --> 00:31:49,440 we have, we have a very sort of free flowing, you know, sharing of information 498 00:31:49,440 --> 00:31:51,150 and what works, what doesn't work. 499 00:31:51,420 --> 00:31:54,060 But then of course, that stays within that specific firm. 500 00:31:54,060 --> 00:31:58,080 And so I think we're seeing different strategies being applied, and I think 501 00:31:58,770 --> 00:32:03,570 the chief AI officer could be, could, could work or not work, probably most 502 00:32:03,570 --> 00:32:06,960 dependent on the person and how much, um, you know, runway you actually 503 00:32:06,960 --> 00:32:08,880 give them how empowered they are. 504 00:32:08,970 --> 00:32:10,080 I think so. 505 00:32:10,170 --> 00:32:10,530 Yeah. 506 00:32:11,280 --> 00:32:12,240 Agreed. 507 00:32:12,660 --> 00:32:17,520 Um, you touched on something that I wanna dive into a little bit, and 508 00:32:17,520 --> 00:32:24,720 that is investment and ROI, um, the law firms, at least here in the US 509 00:32:24,720 --> 00:32:29,850 are primarily partnerships and they mostly operate on a cash basis. 510 00:32:30,480 --> 00:32:30,660 Mm-hmm. 511 00:32:31,050 --> 00:32:34,380 Um, which means that they clear the books at the end of the year, they distribute. 512 00:32:35,055 --> 00:32:40,335 Profits to the partners and they start the year at a zero and do it all over again. 513 00:32:41,145 --> 00:32:46,335 And we are entering into an era where there's gonna need to be capital 514 00:32:46,335 --> 00:32:52,005 investment, short, mid, and long-term capital investment to build out the, 515 00:32:52,035 --> 00:32:56,895 the AI infrastructure because it's gonna be more than just buying lag seats. 516 00:32:56,925 --> 00:32:57,255 Right. 517 00:32:57,255 --> 00:32:58,035 There's, there's. 518 00:32:59,054 --> 00:33:04,004 Um, there is foundational level work at an infrastructure level that needs to happen 519 00:33:04,004 --> 00:33:07,995 to support, um, what AI runs on top of. 520 00:33:08,685 --> 00:33:15,945 So, you know, I've been beating the drum that law firms have historically 521 00:33:16,004 --> 00:33:21,735 over indexed on ROI, and I've been an advocate of quit thinking about 522 00:33:21,735 --> 00:33:25,665 ROI right now because it's gonna be hard to calculate a number. 523 00:33:26,280 --> 00:33:29,969 A spreadsheet because we're, we're in, we're in the midst of change, right? 524 00:33:29,969 --> 00:33:32,669 Like it actually, you may have a negative ROI because you're gonna 525 00:33:32,669 --> 00:33:34,469 reduce billable hours for right now. 526 00:33:34,709 --> 00:33:40,290 But you're learning the, the, the, the r the return in ROI is your learning. 527 00:33:41,010 --> 00:33:45,510 Um, and this is where, this is an era of r and d. Um. 528 00:33:46,560 --> 00:33:50,429 You know, long term there has to be ROI or you can't continue to invest. 529 00:33:50,730 --> 00:33:55,050 But in the short term, I feel like law firms are, are overindexing, not 530 00:33:55,050 --> 00:34:00,719 all of them, but a good percentage are overindexing on what is my ROI right now. 531 00:34:01,679 --> 00:34:02,490 Do you feel that way? 532 00:34:02,490 --> 00:34:07,350 I mean, one, one very clear ROI that we've identified across many of the, of the 533 00:34:07,439 --> 00:34:12,330 firms that we work with is the reduction of work where you're actually not billing. 534 00:34:13,905 --> 00:34:17,325 Which then creates the opportunity to go out and either pitch for more work 535 00:34:17,835 --> 00:34:19,875 or, you know, actually bill for more. 536 00:34:20,385 --> 00:34:23,505 And many of the firms that we work with are, they're at, they're 537 00:34:23,505 --> 00:34:25,545 sort of at the supply constraint. 538 00:34:25,545 --> 00:34:29,295 Like they actually cannot take on more work because they're so fully staffed. 539 00:34:29,925 --> 00:34:36,735 And when, you know, I agree with you that regardless of how you 540 00:34:36,735 --> 00:34:38,175 want to calculate a number, like. 541 00:34:38,730 --> 00:34:42,000 It's kind of go time, like it's here, it's here to stay. 542 00:34:42,270 --> 00:34:43,290 It's undeniable. 543 00:34:43,920 --> 00:34:50,100 However, I, I also do appreciate, you know, the, the ability, the ability 544 00:34:50,100 --> 00:34:51,300 to build a business case around it. 545 00:34:52,110 --> 00:34:56,280 And I think that can come either in the form of sort of, you know, aggregated 546 00:34:56,280 --> 00:35:02,430 numbers or data, but also just the fact that you were able to win that deal. 547 00:35:03,150 --> 00:35:09,090 Because you could tell a story of how your firm leverages AI and how you as a partner 548 00:35:09,090 --> 00:35:14,190 and as a team will leverage AI to get them the best outcome at the best price, right? 549 00:35:14,940 --> 00:35:19,860 And so I think it's no longer a question of do we use this to be faster? 550 00:35:19,860 --> 00:35:21,150 Will we reduce billable hours? 551 00:35:21,150 --> 00:35:26,160 Like if your top firm, you have to use AI to deliver the best 552 00:35:26,160 --> 00:35:30,150 services, the highest quality services at the quickest turnaround. 553 00:35:31,425 --> 00:35:37,935 So there have been studies from McKinsey, Morgan Stanley, Goldman 554 00:35:37,935 --> 00:35:44,625 Sachs, that are estimating 30 to 50% of associate hours being automated away. 555 00:35:45,375 --> 00:35:47,685 And I we're, we're, we're not quite there yet. 556 00:35:47,925 --> 00:35:51,225 But, um, I think we're heading in that, in that direction. 557 00:35:51,795 --> 00:35:52,480 I'm curious about. 558 00:35:53,625 --> 00:35:58,485 With this trajectory that we're on that seems to be moving steeply, 559 00:35:58,695 --> 00:36:00,045 uh, upward into the right. 560 00:36:00,465 --> 00:36:07,515 How long do we have before we're in a place where 30 to 50% of associate 561 00:36:07,515 --> 00:36:12,855 hours are no longer part of the revenue stream for these law firms? 562 00:36:13,215 --> 00:36:17,385 Well, one thing those studies don't take into account is how 563 00:36:17,385 --> 00:36:20,085 much more work there's going to be. 564 00:36:21,285 --> 00:36:24,795 So that's looking at the bucket of work that's available today. 565 00:36:25,395 --> 00:36:29,595 Which part of it do we think that AI can do a, you know, pretty good job of out of 566 00:36:29,595 --> 00:36:35,115 the box or with a system like ro And what they fail to do is they fail to say, okay, 567 00:36:35,595 --> 00:36:38,925 well, because agents will be doing work, 568 00:36:41,805 --> 00:36:43,665 companies will sue more other companies. 569 00:36:44,370 --> 00:36:48,720 There will be more deals, there will be more legal diligence 'cause you're 570 00:36:48,720 --> 00:36:50,160 gonna want to do it from the sell side. 571 00:36:50,160 --> 00:36:52,650 And there's more companies on the buy side who are gonna want to 572 00:36:52,650 --> 00:36:57,990 diligence the company 'cause it's gonna be faster and more accessible. 573 00:36:58,920 --> 00:37:04,170 And so even if there's less associates hours, doing particular 574 00:37:04,170 --> 00:37:08,940 tasks like document review and a diligence, doesn't mean that we'll 575 00:37:08,940 --> 00:37:10,080 necessarily have fewer associates. 576 00:37:11,055 --> 00:37:11,924 Will be more work. 577 00:37:12,315 --> 00:37:16,634 Now, how that gets distributed, where in the value chain, I think is a little 578 00:37:16,634 --> 00:37:22,694 bit sort of TBD, but I think that the associates who are coming out of, of 579 00:37:23,595 --> 00:37:29,565 law school with the skills to leverage AI and who are business-minded and 580 00:37:29,565 --> 00:37:33,075 problem solver and charismatic and can work on clients, sort of, you 581 00:37:33,075 --> 00:37:36,165 know, understand that relationship, those will do extraordinarily well. 582 00:37:36,735 --> 00:37:40,185 And it's also one of the reasons that why we're really excited about our law school 583 00:37:40,185 --> 00:37:41,895 program, which we're launching this week. 584 00:37:42,315 --> 00:37:46,605 So we've partnered up with nine of the leading law schools in the US and not 585 00:37:46,605 --> 00:37:49,694 only are we sort of, you know, getting people access to the system, we're 586 00:37:49,694 --> 00:37:55,095 working with the deans and the law school professors to construct curriculums 587 00:37:55,424 --> 00:38:01,065 that will equip the law school students with the skills to actually bring this 588 00:38:01,065 --> 00:38:02,835 into their workplace once they graduate. 589 00:38:02,835 --> 00:38:07,815 And so we're trying to take responsibility for the full, you 590 00:38:07,815 --> 00:38:09,375 know, in Swedish we would say from. 591 00:38:11,025 --> 00:38:15,254 From sort of corn to bread, but sort of all the way across the value chain. 592 00:38:16,424 --> 00:38:16,995 Interesting. 593 00:38:17,444 --> 00:38:18,345 What are 594 00:38:20,595 --> 00:38:25,365 three things that AI is really good at in legal work, and what are three, three 595 00:38:25,665 --> 00:38:28,395 things that it's not good at today? 596 00:38:30,915 --> 00:38:32,805 AI is really good at reading. 597 00:38:33,105 --> 00:38:37,035 It's really good at finding things in documents, contracts, 598 00:38:37,335 --> 00:38:38,085 whatever it might be. 599 00:38:38,640 --> 00:38:40,379 That it's extraction is really good. 600 00:38:41,940 --> 00:38:46,440 It's really good at looking at a lot of data, big buckets of data 601 00:38:46,529 --> 00:38:52,049 and finding the right thing, sort of needled in a haystack, a bit dependent 602 00:38:52,049 --> 00:38:53,279 on how you structure your data. 603 00:38:53,399 --> 00:38:55,290 But AI is really good at that. 604 00:38:56,399 --> 00:38:58,740 I think AI is really good at ideation. 605 00:38:59,040 --> 00:39:04,560 He's actually really good at, um, finding ideas and seeing patterns that 606 00:39:04,560 --> 00:39:07,589 aren't necessarily as natural to humans. 607 00:39:08,310 --> 00:39:13,830 So ideating on a case strategy or on how you might be able to rephrase a paragraph 608 00:39:13,830 --> 00:39:16,410 or a clause to actually get that through. 609 00:39:17,160 --> 00:39:23,759 Extraordinarily good at what AI maybe isn't as good as today is. 610 00:39:25,800 --> 00:39:33,000 When the context overflows and you, for instance, need to understand how a 611 00:39:33,000 --> 00:39:39,060 very complicated legal research query, um, should be answered using a ton of 612 00:39:39,060 --> 00:39:44,400 different, um, you know, case law and precedent, uh, it's difficult for the 613 00:39:44,400 --> 00:39:49,230 agents to actually look at all of that and then come up with a definitive answer 614 00:39:49,680 --> 00:39:52,320 because they don't necessarily understand how everything sort of ties together. 615 00:39:54,620 --> 00:40:00,375 I think AI is not very good at very complex drafting, right? 616 00:40:00,375 --> 00:40:05,384 Like if you needed to draft a full SPA from scratch, it'll not do a good job. 617 00:40:05,835 --> 00:40:10,905 However, if you provide it a precedent in context, it does a really good job. 618 00:40:11,595 --> 00:40:13,965 So again, it sort of goes back to is AI good out of the box? 619 00:40:14,400 --> 00:40:18,600 Or is AI good within an ecosystem like leg where it can actually perform that 620 00:40:18,600 --> 00:40:22,680 task Well, so I think the future of software is a very integrated one. 621 00:40:23,190 --> 00:40:28,290 It's where you are able to launch agents, give them tasks, give them 622 00:40:28,290 --> 00:40:32,790 commands, and they're going to be able to work with the rest of your ecosystem 623 00:40:32,790 --> 00:40:34,710 and your stack in order to do that. 624 00:40:35,460 --> 00:40:40,529 And so for legal research specifically, one of the reasons why, um, I think that 625 00:40:40,680 --> 00:40:43,440 task has been quite hard is because. 626 00:40:45,780 --> 00:40:51,210 AI always gave the correct answer on research query that 627 00:40:51,210 --> 00:40:52,890 would sort be profoundly. 628 00:40:53,685 --> 00:40:57,645 Scary thing like that means that it's always correct no matter what. 629 00:40:58,155 --> 00:41:01,725 And the reason why it's hard to make it correct all the time for every type 630 00:41:01,725 --> 00:41:06,435 of query is that the problem space is so large because you can ask so 631 00:41:06,435 --> 00:41:07,635 many different types of questions. 632 00:41:08,235 --> 00:41:11,355 And what's important is, of course, one, the underlying data that you 633 00:41:11,355 --> 00:41:14,715 work with, but two, how the data is structured and how you actually 634 00:41:14,715 --> 00:41:16,135 let an agent traverse that data. 635 00:41:17,130 --> 00:41:18,210 And now how you make sense of it. 636 00:41:18,330 --> 00:41:21,810 You know, you need a Tator, you need everything to sort of line up to this. 637 00:41:22,470 --> 00:41:26,280 And we've invested a lot of resources in our US legal research. 638 00:41:26,430 --> 00:41:31,320 And so with both state and federal case law, um, we're now starting to 639 00:41:31,320 --> 00:41:35,730 work more and more with how all of this can be tied together in a very 640 00:41:35,730 --> 00:41:41,310 cohesive, um, experience, both for the user and for the agents that are 641 00:41:41,310 --> 00:41:42,600 then leveraging all of this together. 642 00:41:43,620 --> 00:41:44,400 That makes sense. 643 00:41:44,880 --> 00:41:45,960 So, um. 644 00:41:46,605 --> 00:41:50,205 We only have a few minutes left, but one area I wanted to touch on 645 00:41:50,205 --> 00:41:54,225 that I know it's, IM, uh, important to you is like collaboration. 646 00:41:54,225 --> 00:41:54,285 Yeah. 647 00:41:54,765 --> 00:41:58,185 So Info Dash is a legal collaboration platform. 648 00:41:58,605 --> 00:41:58,964 Yes. 649 00:41:58,964 --> 00:42:05,205 We do not compete with Lara, um, or Harvey or any other practice of law tool. 650 00:42:05,205 --> 00:42:05,805 We really. 651 00:42:06,900 --> 00:42:11,850 Um, we are operate in the compliment, like I think the practice of law. 652 00:42:11,880 --> 00:42:17,430 If you were to draw a pie chart of all of the use cases that fall into 653 00:42:17,940 --> 00:42:21,270 the collaboration bucket, how law firm clients want to collaborate 654 00:42:21,270 --> 00:42:24,360 with their law firms, um, the. 655 00:42:25,995 --> 00:42:29,325 The practice of law represents a fairly narrow slice today. 656 00:42:29,535 --> 00:42:32,745 I think that slice will grow wider as we become tech enabled. 657 00:42:33,404 --> 00:42:39,345 But there are things like, you know, um, matter intelligence, so legal 658 00:42:39,345 --> 00:42:45,254 team information, um, you know, legal project management tasks, uh, 659 00:42:45,285 --> 00:42:47,895 financial information reporting. 660 00:42:47,984 --> 00:42:48,734 Um. 661 00:42:49,904 --> 00:42:52,904 You know, knowledge management, there's all these other things outside. 662 00:42:53,145 --> 00:42:57,345 You guys announced your portal product like two days before TLTF summit 663 00:42:57,765 --> 00:42:58,904 and everybody's coming up to me. 664 00:42:58,904 --> 00:42:59,955 They're like, are you worried about la? 665 00:43:00,525 --> 00:43:05,174 I'm like, you clearly don't understand what a legal extranet platform does. 666 00:43:05,505 --> 00:43:10,725 Yes, law firms will want to collaborate on work product with their customers, 667 00:43:10,725 --> 00:43:11,700 and I think that's where you guys. 668 00:43:12,645 --> 00:43:15,524 Play, but there's so, so much more to that. 669 00:43:15,524 --> 00:43:19,785 But tell me your, like long-term vision about how you see Lara 670 00:43:19,785 --> 00:43:21,464 playing in that collaboration space. 671 00:43:22,484 --> 00:43:26,205 Yeah, so I think it's, it's really sort of two things at the same time, um, we're 672 00:43:26,205 --> 00:43:31,634 working with firms who are looking for ways to scale their delivery method. 673 00:43:32,325 --> 00:43:33,944 Take Debevoise as an example. 674 00:43:34,545 --> 00:43:38,205 Debevoise and the partner Avi there just launched their star 675 00:43:38,205 --> 00:43:41,399 portal on Laua and they're working with clients like Blackstone. 676 00:43:42,630 --> 00:43:45,330 They're also working with other private equity firms, and they're 677 00:43:45,330 --> 00:43:50,490 allowing all those private equity firms to access structured workflows, 678 00:43:50,550 --> 00:43:57,300 work products, databases, and content that gets them value faster. 679 00:43:58,110 --> 00:44:00,330 And so what they're, what they're doing is they're, they're 680 00:44:00,360 --> 00:44:01,920 scaling their delivery method. 681 00:44:03,150 --> 00:44:06,210 And I think that's very interesting because that's a way to actually 682 00:44:06,210 --> 00:44:11,130 move away from, you know, how you're actually leveraging the billable hour. 683 00:44:11,910 --> 00:44:17,640 And I think the real pioneers and the client excitement is real. 684 00:44:18,209 --> 00:44:25,350 They really value the progressiveness of this. 685 00:44:26,399 --> 00:44:29,399 On the other hand, you have, um, enterprises. 686 00:44:31,379 --> 00:44:34,770 That are working with many firms who want to consolidate the 687 00:44:34,770 --> 00:44:35,879 way that that actually works. 688 00:44:35,970 --> 00:44:40,230 So instead of having, you know, 20 panel firms and the way that you 689 00:44:40,230 --> 00:44:44,460 manage all of that work, um, they want to be able to work jointly 690 00:44:44,730 --> 00:44:46,470 on specific matters with them. 691 00:44:47,160 --> 00:44:47,580 And so. 692 00:44:48,750 --> 00:44:51,900 One of the things that we experienced as part of our fundraise 693 00:44:51,960 --> 00:44:55,290 is that collaborating with our law firm is quite inefficient. 694 00:44:55,800 --> 00:45:01,710 It's all on email and when we have, um, to-dos or when we need to communicate 695 00:45:01,710 --> 00:45:06,030 or when I need to go and read our shareholders agreement, it's great 696 00:45:06,030 --> 00:45:11,820 if that's already sort of accessible to me, and then multiple clients of 697 00:45:11,820 --> 00:45:13,500 theirs have the same type of problems. 698 00:45:13,710 --> 00:45:16,110 So if AI be doing those tasks. 699 00:45:18,120 --> 00:45:20,580 You know, they might as well deliver them via AI and via 700 00:45:20,580 --> 00:45:21,839 software to all those clients. 701 00:45:22,410 --> 00:45:23,520 Yeah, that makes sense. 702 00:45:23,700 --> 00:45:24,930 We're actually excited. 703 00:45:25,020 --> 00:45:29,310 Um, I mean, Harvey's announced a, a similar like client facing, you 704 00:45:29,310 --> 00:45:31,950 know, again, we get asked a lot, Hey, are you worried about these 705 00:45:32,009 --> 00:45:37,290 well-funded players coming in and talking about legal collaboration? 706 00:45:37,710 --> 00:45:42,150 And we're ex actually excited about it because it has been a very historically. 707 00:45:43,080 --> 00:45:48,810 Neglected space, like, I mean, the dominant player there is hiq that was 708 00:45:48,810 --> 00:45:50,850 bought by Thomson Reuters in 2019. 709 00:45:51,120 --> 00:45:55,710 The product hasn't changed a whole lot in, you know, the last seven years and 710 00:45:55,740 --> 00:45:59,640 there's only been one player in town and they own half the market share. 711 00:46:00,150 --> 00:46:05,070 So I, you know, I like to take an abundance mindset to thinking 712 00:46:05,070 --> 00:46:08,400 through this and I think it, it will create opportunity. 713 00:46:08,400 --> 00:46:11,880 Maybe at some point info dash will have integrations with. 714 00:46:13,155 --> 00:46:15,525 Lara and other AI platforms. 715 00:46:16,095 --> 00:46:19,485 Um, so yeah, I'm an, I'm an optimist. 716 00:46:19,485 --> 00:46:20,115 I think that. 717 00:46:21,430 --> 00:46:24,735 I think it's been a, a, a place that's been historically neglected. 718 00:46:25,125 --> 00:46:26,415 I think that law firms. 719 00:46:26,715 --> 00:46:31,005 And then, uh, this is the last question I'll ask you 'cause um, I know we're 720 00:46:31,005 --> 00:46:35,055 running over, but I think that law firm clients are gonna want different 721 00:46:35,085 --> 00:46:36,735 information from their law firms. 722 00:46:37,125 --> 00:46:43,300 So when you can turn around documents in hours or sometimes minutes, like 723 00:46:43,300 --> 00:46:48,825 you're gonna start measuring your law firm partners based on turn time on. 724 00:46:49,455 --> 00:46:55,425 Number of defects, um, you know, volume there, it, whereas if it takes a week 725 00:46:55,425 --> 00:47:00,345 like that, that's too slow to have a dashboard around how quickly legal 726 00:47:00,345 --> 00:47:02,085 processes are being turned around. 727 00:47:02,295 --> 00:47:06,495 So we think that there, there will be an appetite for. 728 00:47:07,575 --> 00:47:11,985 Inside legal teams to see metrics around these sorts of things. 729 00:47:11,985 --> 00:47:13,695 I don't do, do you share that view? 730 00:47:13,905 --> 00:47:13,965 Yeah. 731 00:47:13,965 --> 00:47:17,535 Well, I think like there's different categories of work and uh, like probably 732 00:47:17,535 --> 00:47:19,215 some, some stuff fit within that. 733 00:47:19,215 --> 00:47:21,705 And then there's gonna be very large complex matters. 734 00:47:21,705 --> 00:47:25,335 We're gonna be working for a longer time and still it's gonna be very 735 00:47:25,335 --> 00:47:29,055 useful to collaborate, um, with the enterprise or the customer that 736 00:47:29,055 --> 00:47:30,645 you're working with and the firm. 737 00:47:31,005 --> 00:47:33,495 And I think. 738 00:47:35,190 --> 00:47:39,150 It's the most exciting time in history to be a lawyer because, and a legal 739 00:47:39,150 --> 00:47:45,029 professional because no lo you know, never have there been so many talented 740 00:47:45,029 --> 00:47:48,240 people around the world working on building delightful software. 741 00:47:48,540 --> 00:47:53,670 And the pace of that development is staggering and a lot of fun. 742 00:47:53,940 --> 00:47:59,549 And just as you said, I'm also very, um, optimistic about the future world 743 00:47:59,549 --> 00:48:02,370 of abundance and how everybody's going to collaborate within its. 744 00:48:02,790 --> 00:48:03,210 Yeah. 745 00:48:03,629 --> 00:48:03,899 Yeah. 746 00:48:03,899 --> 00:48:11,430 It's easy to, you know, there have been lots of revolutions and transformations 747 00:48:12,029 --> 00:48:17,009 throughout human history, and going into them is always scary because 748 00:48:17,009 --> 00:48:19,710 change is hard, but I, I'm with you. 749 00:48:20,535 --> 00:48:23,895 People ask all the time, are you worried about somebody vibe, coding info dash? 750 00:48:24,315 --> 00:48:30,345 Which makes me chuckle because, you know, if they only knew how many, like how much 751 00:48:30,345 --> 00:48:34,245 learning it took, it's not the writing the code piece, that's the easy part. 752 00:48:34,635 --> 00:48:37,785 It's the learning, oh, you know what, that's the wrong API because 753 00:48:38,055 --> 00:48:41,985 that one gives you something that's different than what you actually need. 754 00:48:41,985 --> 00:48:43,605 I need this API over here. 755 00:48:44,115 --> 00:48:47,955 Um, so yeah, I, I, I think we're. 756 00:48:48,600 --> 00:48:54,840 We're headed for an era of abundance and I take much, I lean much more on the 757 00:48:54,840 --> 00:49:00,120 utopian, um, side than the dystopian, but I do think it's gonna be a bumpy ride. 758 00:49:00,509 --> 00:49:02,190 It's not gonna be smooth sailing. 759 00:49:02,430 --> 00:49:02,610 Yeah. 760 00:49:03,120 --> 00:49:07,380 Change is hard and um, I think there will be some bumps along 761 00:49:07,380 --> 00:49:08,490 the way, but we'll figure him out. 762 00:49:10,530 --> 00:49:10,945 I think so too. 763 00:49:11,835 --> 00:49:12,404 Awesome. 764 00:49:12,645 --> 00:49:14,265 Well, this has been a great conversation. 765 00:49:14,265 --> 00:49:17,595 I really appreciate you taking the time and sticking with me. 766 00:49:17,595 --> 00:49:20,265 We've been trying to get this on the books for at least a year. 767 00:49:20,564 --> 00:49:21,794 We finally got it done. 768 00:49:21,855 --> 00:49:25,814 Um, so I, I really appreciate you, uh, spending some time today. 769 00:49:27,075 --> 00:49:27,379 Thank you so much. 770 00:49:27,379 --> 00:49:28,665 Uh, that was fantastic. 771 00:49:28,875 --> 00:49:29,660 All right, well chat soon. 772 00:49:31,770 --> 00:49:32,160 Take care. 773 00:49:32,279 --> 00:49:34,560 Thanks for listening to Legal Innovation Spotlight. 774 00:49:35,100 --> 00:49:38,580 If you found value in this chat, hit the subscribe button to be notified 775 00:49:38,580 --> 00:49:40,080 when we release new episodes. 776 00:49:40,589 --> 00:49:43,259 We'd also really appreciate it if you could take a moment to rate 777 00:49:43,259 --> 00:49:45,899 us and leave us a review wherever you're listening right now. 778 00:49:46,470 --> 00:49:49,200 Your feedback helps us provide you with top-notch content. -->