How Solve Intelligence Is Building the Future of AI Patent Law

Diana Hu 0:00 So we're gonna welcome here Chris, Angus, and Sanj, who just were here not too long ago. They started this company called Solve Intelligence in 2023 and went through YC in that summer of 2023. Chris is the CEO, Angus is the CTO, and Sanj is the chief research officer. And what Solve does is makes AI software for patent attorneys, and more than 700 IP teams use you today. This is just two and a half years later, which is incredible. And law- big law firms use you, like Perkins Coie, DLA Piper, and in-house teams at big companies like Intel and Siemens. And to this date, they've raised $55 million in total. And not too long ago, they were in your shoes, just 2023, you're new grads, and we're gonna go through that. I mean, tell us the story of how you started, because Chris, you had just finished your PhD here at UCL. Your-- And you two dropped out from quant jobs to do Solve. Tell us about the beginning.

Chris 1:14 [clears throat] Yeah. Um, so I was doing a PhD in AI at, at UCL. Um, finished that a little bit early, so I had a few months to spare and, um, started a side project with my brother Angus here, building AI software for patents. The way that we kind of got into it initially was, one, that we'd been on the inventor side of things. One thing about patents is that they involve the inventor a lot because they're very technical. And so coming from the research background, we'd been on that side of things. We were familiar with some of the time and cost pain points. That was one thing. And the other thing is that my girlfriend Yasmin happens to be a patent attorney, so I was familiar with the fact that there wasn't a lot of software available. And of course, in 2023, we just had this big step change in the performance of large language models. And so Angus and I got interested in doing this as a, as a side project. Um, but I wasn't seriously considering or we weren't kind of seriously considering doing a startup. Uh, on the back of that, we-- my plan was to go off and do AI research in Singapore. I thought it'd be cool to go and live somewhere else, and so had a job offer to go and, and do that, and that was my plan. Um, but then I got talking to, to Sanj, um, and you know, Sanj can talk about why he decided to do this, but he was going all in on, on AI, super excited about it, and I was just asking for career advice, basically. And then one thing led to another. We started hacking away on what became, um, what became Solve, and we got super excited about it. Went all in, um, got into Y Combinator, and then, um, yeah, built since then.

Diana Hu 2:39 Yeah, you've been in this huge trajectory just two years and a-- two years and a half later. And one actually thing that many people may have not know about Solve, this is a two-year-old startup. When you applied to YC, it was basically, like, a two-week-old idea. You just had graduated, and there was not much built. It was basically just an idea on paper. What do you think were your chances of getting into YC?

Chris 3:05 Uh, very small. Very small. I mean, we basically, um... So wh- when I, when I had this, um, call with Sanj, I, as I say, I was asking about career advice, and I shared screen 'cause we just got talking about what he was building, which sounded super interesting, and then I started talking about what Angus and I were building. Shared screen. He got, um, super excited about what we were building, so I went round to his house the next day, and we started building together at h- his apartment. And then Sanj, I think you were the one who said that the YC deadline's in a week or something, right? And, um-

Diana Hu 3:35 So you applied in a whim.

Chris 3:36 Yeah. Yeah. Well, we sort of thought, "Well, we may as well go through the process of filling out the questions." And so we, we kind of did that throughout that week, and, um, that was actually the first time that we started talking to actual patent attorneys, to the customers, which of course we should have done at the beginning. But filling in the YC form was super useful for us, actually kind of really formulating what it is that we're trying to do, and we ended up sort of pivoting the idea a little bit, still around the patent space. But by the end of the week, through the process of talking to patent attorneys, then continuing to build based on what they were telling us and, and interestingly filling in the form to then submit it, I think on the deadline or the day before, I can't remember. Um, by the time we did that process, I was fully convinced, and I think these guys were as well, that this was gonna be a big thing that we should go all in on. So turned down the job offer, went, went all in on the startup, and then, um, and then, yeah, we ended up getting into YC, but we did think it was a, a long shot when we submitted the application.

Diana Hu 4:27 And-

Chris 4:27 Yeah.

Diana Hu 4:27 And Angus, you actually left a quant job to do this. W- uh, you didn't have much and decided to take the leap of faith.

Speaker 4:34 Yes. So, um, if I think back to when I was working the quant job, I think it was, if I'm honest my- with myself, it was basically salary.

Diana Hu 4:43 Yeah.

Speaker 4:43 Um, which is probably true for a lot of people here as well. It was always my dream to start a company, and it's definitely much more satisfying work, kind of speaking to users, telling, you know, them telling you they hate your product or they love your product. It's just much more satisfying work, and so I'm very glad that we did it. Chris and I were always massive YC fanboys, um, and so it felt like something that was kind of way off, like, you know, not thinking about it. So thankfully, Sanj got us to apply. Um, and as Chris said, doing the process of applying, we pivoted the entire business. So yeah.

Sanj 5:14 Yeah, I mean, for me, my big motivation was kind of seeing DeepMind spring out of the PhD lab. 'Cause I previously did a PhD, then did a quant job. And so I'd seen kind of Demis or whatever. I'd heard stories about how Demis had come out of my lab and founded DeepMind, and so poached the name Solve Intelligence 'cause I thought, you know, like you said, it's more exciting to build something very ambitious and dream big and think, "Oh, how can I solve, like, really fundamental problems?" than kind of start with something very, you know, niche and micro. And then, yeah, met Chris and Angus and was, like, phenomenally impressed. And I think that's a thing when you find co-founders where you're just really impressed. You shouldn't hesitate. You should just go for it.

Diana Hu 5:57 And one of the wildest thing now reflecting of this, Sanj, is actually your company right now with Solve Intelligence is actually bigger than when DeepMind was. ... got sold. You have way more employees than when they got sold and actually have made more revenue than them. Can you believe that?

Speaker 6:13 It's surreal. I think-

Diana Hu 6:14 Than, than Demis. I know he's like a hero here. [laughs]

Speaker 6:17 Yeah, I think it's surreal, and I think there's also the channeling the very practical vibe that we also have. I think there's, like, a really nice balance between, okay, we have, like, really deep, smart thinkers at Solve, but also just incredibly practical customer-focused people. And because we were able to get revenue pretty early, like, straight away, basically, after joining YC, it gives you a lot of freedom because then you can say, "Look, we're a company on our own two feet." And I think the issue for DeepMind, of course, I didn't... I agree with you. They, they shouldn't have sold, is, like, they, they didn't have revenue at the time, and that then kind of does put you in a position where you are having to, like, look at different options.

Diana Hu 6:55 Yeah. And one interesting thing about all of you, none of you come from a law background. You're all hardcore engineers, PhDs, and scientists. Why patent? That it seem... I mean, you mentioned your, your, your girlfriend, right?

Chris 7:10 Yeah. I mean, we just thought, um... We, we, we really wanted to get involved in the... You're kind of getting on this sort of AI cycle that's happening. We just thought there were so many problems that had been unlocked by the step change in the performance of the models that previously weren't solvable and potentially now might be solvable. Um, turns out, there's j- I mean, as everyone knows, there's just now loads of applications that you can build. But, um, we all came from technical backgrounds, research backgrounds. We're familiar with patents. It seemed like an interesting area to go into because they are kind of very, very technical, and it's a big industry. There's over $300 billion spent every year across the patent process, and yet there's hardly any technology being used. So it's a combination of interesting technical problem, greenfield space, a lot of demand and, um, big opportunity.

Diana Hu 7:56 Yeah.

Speaker 7:56 I think in... I was gonna say, I think in a weird way, it kind of helped our product a little bit because we had no opinions on how a patent attorney should work or how the product should be, and so we basically had to speak to customers, which you should do anyway. Um, but we had no choice.

Chris 8:11 Yeah. One, one, just to add one more thing on that is when, when we started talking to customers, a lot of them would say, um, there's no way that AI is gonna be able to help with the stuff that they were doing, drafting patent applications and so on, because the work's very complicated. You're having to deal with these sort of, um, multimodal inputs and outputs, like chemical structures, technical drawings, and so on. And especially back in 2023, this seemed like a very sort of, like, um, that, that this was not gonna be possible by AI. And I think having the naivety that we had at the time, we thought, well, there's no reason why it, it should be like that, especially with the, you know, trajectory that the models were on. And so it did help us, I think, a little bit coming in with a completely blank slate.

Diana Hu 8:48 There was something about, uh, about you moving to San Francisco during the batch that got you to raise your level of execution and ambition, because within two weeks, you basically... Of the start of the batch, you actually got your first paid customer and shipped your product. I think if, uh... And you kept going on that rate. What, what was that catalyst of being in San Francisco that helped you move a lot faster?

Chris 9:11 Yeah. So firstly, one thing I... I mean, we obviously decided to move back to London after San Francisco, but I massively agree that the level of ambition in San Francisco is like nothing anywhere else, but especially, uh, here in Europe. I think, um, even... As soon as you get there, you know, you come off the plane in San Francisco out of the airport, and, um, you get in your Uber, and as you're driving in, you see all the, all the ads are for sort of B2B AI companies or Nvidia GPUs. Um, you go into San Francisco, there's Waymo's driving around, self-driving cars. You go to any kind of social event or startup event, everyone there is in tech. They're building startups. They're in exciting companies and so on. It's just totally different to everywhere else. So I fully agree that if you guys are interested in tech, which clearly you are, you should definitely go to San Francisco, and you should meet people there. And, um, you'll... you will notice that the level of ambition is totally different to, to here in London. But in terms of launching quickly, I mean, like, um, Sanj, San- San said, we, you know, we, we really are sort of, uh, uh, we, we drank the Kool-Aid with Y Combinator, and, um, one of the things that YC recommended was to launch early. And so I think within the first one or two weeks of the batch, um, we decided to launch. And, uh, um, you know, my recollection of it was, um, not to get too graphic, but I just came out the shower with a towel on, and I said to Sanj and Angus on the Sunday, "I reckon we should launch, and we're going to the YC retreat the next day." And they were like, "We have no product. We have no product to actually ship." But I... You know, I think we then realized that if we launched tomorrow, it would be a real forcing function to actually get the product shipped, get it hosted, get something that we could actually sell and provide to the customers. So we did do that. Made the first sale on the Friday that week. We had no product, so we then spent the following weekend getting actually spun up and serving it on the Monday. Um, he churned the following week, but at least we got the good feed- we got the feedback to improve it. But yeah.

Diana Hu 10:54 That's pretty good.

Speaker 10:55 I would say also during the batch, the group office hours were quite a good forcing function 'cause you basically sit in a group with similar companies all at a similar stage, and you kind of say, "In two weeks, my revenue's gonna be this." Um, and you kind of want to fail that goal 50% of the time. And, um, you have this sort of, like, public shame if you turn up the next two weeks and you haven't hit your revenue goal, and it really helps. And I think it's basically we've always set goals throughout the history of the company, and so I think it works very well.

Diana Hu 11:19 And I guess another question with during the batch is you started to really... None of you have any sales background, right?

Speaker 11:28 No.

Diana Hu 11:28 And at some point, you got to several millions in annual revenue it- without any salesperson. It was just, just you guys, the founders. How did you do that? I mean, in other... Sometimes people say, "Oh, we need to hire a salesperson," but you figured it out.

Chris 11:42 Yeah. So yeah, none of us had-- I'd never sold anything in my life. I just spent my whole life at university kind of doing research and academia and so on. Um, never really had a proper job other than just sort of internships and then went straight into doing this. So I really had, you know, we really had no idea what we're doing. But, um, you know, it-- there, there was no sort of complicated like hack or anything. We basically launched super early, um, sold the product. We didn't do free trials, so the only feedback that we were getting was from real paying customers. And, um, at the time at least, that was extremely useful because we knew nothing about patents in hindsight, and so we needed a real signal to optimize for, which paying customers is much more valuable than the, than the free trials. And so we just sort of, you know, plotted growth every month and plotted churn every month. You could see the churn coming down and down as that did, then the growth kind of compounded because customers start expanding, they start paying more, they start saying lots of positive things about your product. You start getting more and more inbound. And so, um, in hindsight, we should have hired much, much earlier, but we didn't know that at the time. We'd never-- I mean, hiring is another sort of thing that we didn't know much about when we started the company. Um, but we, we didn't need to hire in order to get to millions of, of revenue. We could have hired and I think grown faster, but we didn't need to get there. It was really about focusing on product and customer feedback.

Diana Hu 12:57 A question for Sanj. I mean, for a lot of the audience might be curious about why wouldn't a lawyer just use ChatGPT on top of this? Maybe some people think it's like, oh, patent law could be just done in ChatGPT. What, what was the insight that you guys figured out?

Sanj 13:14 I think when we even started the company, we liked the idea of patents because it was like a small, like very focused idea, but it had so much room to grow. Um, if you think about what it means to write a patent, you actually have to deeply understand the technology so that you can then patent something that's novel and non-obvious. And then the reward signal, um, which is like whether the patent got accepted, could be like a year and a half way, and maybe even litigation is, you know, five, 10 years down the line. For example, like we know famous YC companies that their company went bust and then the founders ended up multi-millionaires purely from the value of their patents. Like they sued Meta because they had, you know, patents basically. And so bringing it back into like what we do well, I think is we deeply understand the problem. We have customers who want the product to work for them. So it's an amazing experience to have customers basically tell you what they need to build or what they need as in terms of output, and then kind of now of course, scaling evals, scaling data sources, and just really just being ruthlessly efficient, almost like a, an elite hedge fund. Uh, like, um, making sure that you really are like getting the correct, you know, novel idea for this customer who ultimately wants, wants their patent to be valuable.

Diana Hu 14:32 And I think there was another special thing that you guys figured out because of your training with your PhDs at Imperial, UCL. You figure out and solve some really hard technical problems as well, and you were able to match that up and shape it into a-

Chris 14:45 Yeah

Diana Hu 14:45 ... building for patents. What were-

Chris 14:46 Yeah. So yeah, I mean, but pa-patents, it kind of gets categorized as, as legal tech, but it's actually super different to other areas of legal like stuff where you're doing M&A, um, activities or contract review and so on. Um, patents are deeply technical, like Sanj mentioned. You have to involve lots of different multimodal inputs and outputs, so things like, um, chemical structures, technical drawings, biological sequences. You have to handle video data, equations, raw tabular data. Um, and you know, the actual workflows that you go through when you're drafting a patent application or analyzing a patent portfolio, building out these things called claim charts, drafting office action responses, the different kinds of sources that you're having to integrate with, and therefore all the kinds of harnesses in the UX and the workflows you have to build are very, very different, definitely from, um, something that's generic, like a sort of ChatGPT or Claude, but also from some of the more general, uh, legal tools which are good for other areas, but less good for, for patent workflows. And so I think it's just a great niche where it's a big industry, um, but also it requires a lot of technical, um, customized technical stuff to actually get the agents working properly.

Diana Hu 15:50 Now, one of the impressive things for you guys, last year you guys grew 10X in one year. In December, you guys were just 12 and just announced your Series B raise for forty million. And now fast-forward about seven, eight, nine months later, you are already seventy employees and you have multiple offices in London, New York and Munich, which is to the point of now you guys have a bigger company than DeepMind when it got sold, right?

Chris 16:20 Mm-hmm.

Diana Hu 16:21 I think DeepMind was less than-

Sanj 16:22 Yeah

Diana Hu 16:22 ... it was way fewer employees than you guys have right now, right? So don't sell.

Chris 16:26 [laughs]

Diana Hu 16:27 Uh, to the point of being here. So you guys have an office here in London. Tell us about what has been like this, this growth. I mean, you, you've been riding this wave, like what Gustaf mentioned, like this is like a very good time to build and start a company and you're part of that wave. What, what has it felt like?

Chris 16:44 Um, yeah, I do feel like, uh, I think like we all feel like this, but now just feels like a super unique moment in time where there's so much technology change happening across all of these sort of trillion-dollar industries and, um, I feel a lot of sort of like pressure to move super fast. And, and it's very exciting. If you love building products, right now is super exciting. So, um, anyone that's working in finance or a big technology company, I strongly encourage you to, uh, come and join Solve. That's the best option. Second best option, go and, uh, start a company.

Diana Hu 17:16 Which we, we're great. If you're not ready to do a startup-

Chris 17:18 [laughs]

Diana Hu 17:19 ... join our early stage startup like Solve.

Chris 17:21 Um, but yeah, no, it's been, it's been super exciting. I mean, um, the amount of stuff to... I think a good heuristic that I think that you guys mentioned that I certainly agree with is if, if you-- the way that you should not only decide what you're working on, but also how you build your products, is ask yourself, do you feel genuinely excited about the next model release, or do you feel fear of being, like, eaten up by Claude or, um, or GPT?

Chris 17:42 And-

Diana Hu 17:42 How do you guys feel about that with the next, uh, Claude or Fable?

Chris 17:45 So we feel very excited because we feel we're working at the absolute frontier of what these models can do. And, um, you know, I still feel like even if the model performance froze, there's like at least another sort of year of like a lot of agentic stuff that we could be building on to provide value on top of the models. But each new model release unlocks completely new capabilities. We're building on top of all of the models inside our products. We have Gemini, Astra, we've got lots of different open source models. But-- and they all perform differently at different things in the, um, in the patent space. It's a very kind of jagged frontier of, of performance. And so we get very excited with the new model releases. Um, uh, but yeah, it's an exciting time growth-wise and product-wise.

Speaker 18:24 One of the cool things about building an AI company at the moment is just because the models are improving so quickly, you have to kind of rework your entire product and UX every sort of three or four months to, to sort of live in the new age. So when we first started the company, it was very much kind of granular tab autocomplete, and now it's much more agentic, and I think that's gonna carry on happening with just new stuff you can do, basically.

Sanj 18:45 Yeah. And the one thing I'd add that I miss, guess is, I guess is critical on is a lot of the organization level sharing, a lot of the stuff that like, you know, traditional stuff that just makes your product in-incredibly sticky. It's trying to get that foothold in where people just feel confident to dump all their shit, I suppose, inside your product. And so we're like drastically encouraging all of our like major firms, you know, custom deployments, dump all your stuff in as much as possible. We'll work with you to optimize how we can get the most out of your, you know, data where possible. Of course, a lot of confidentiality there, but also your setup, your ways in which you share things. So, um, there's a lot of tricks in the startup playbook that make you sticky.

Diana Hu 19:26 And another interesting thing about you guys is you apply to YC as a London/San Francisco company, and you move to San Francisco for the batch, and then you move back to London. And you-- most of your customers are in the US, and you have an office in New York. What would you tell a founder in this room on how to navigate where to be, especially with visa constraints and customer, where they're gonna be, engineers? How, how do you think about that?

Chris 19:53 So my, my opinion is that everyone who's interested in building startups and technology and so on should definitely go to San Francisco. Um, uh, it's the, it's the most ambitious place. You're surrounded by the most, uh, talented people working on interesting things. But, uh, I don't believe that you have to be in San Francisco to build a generational technology company. In fact, I think there's, um, significant advantages to, uh, to being here in London. I think one of the things in London is that unlike San Francisco, there's actually not that many really good startups. Um, and there's a con to that, which is that if you allow yourself to, you know, compare yourself to the startups around you, then you kind of revert to this mean, where the mean is not very good. Um, but if you ignore that, if you kind of have this sort of siege mentality, then what you have is a bunch of untapped, super technical and smart and hardworking people, and I think that's what London has a huge amount of. And so I think we've benefited a lot from that and the technical team that we've been able to build at Solve Intelligence, more than we're in-- if we were in San Francisco. But having said that, you know, the US is our biggest market. It's our main priority. It's where the majority of our revenue comes from. We sell across six continents. We have a global presence, but you know, I'm basically constantly on a plane. I think in the last, uh, thirty or forty days, I've probably spent two nights in my own bed. Everything else has been going across all these different countries and so on. And so, um, that's just part, that's just part of what's involved. You have to get on a plane, you have to go and see customers, you have to talk to them and so on. Um, but yeah, I don't think you have to stay in San Francisco, but I certainly recommend that you go out, you go out there.

Diana Hu 21:20 So if I summarize, part of it is what was helpful for you guys is you got a lot of the culture of San Francisco installed into your startup culture, and you got that, and then you could start that and continue executing like a San Francisco company and be here.

Chris 21:33 That's right. Yeah. So like Angus said, the, the structure of doing the YC batch where you've got ten weeks, whatever it is, until demo day, and every two weeks you're coming, and you're giving updates, and everyone else is giving updates, and they're all kind of crushing you on their metrics and so on, and you, you sort of have to be more and more ambitious if you're hitting-- if you actually are hitting your goals, I think is an extremely good way to spark your company at a time when you have the least direction because you have zero customers, you have zero pro-products, and that's a very good sort of focusing mechanism. And if you can sort of bring that back and all the ambition and you can kind of continue with that momentum, uh, then it can be pretty powerful.

Diana Hu 22:06 Now, before we open up for a couple questions, we're gonna ask this question for each of you. So half of this room works at a big lab, big tech, quant firms like all of you. Basically like a version of all of you, a younger self two years ago. Uh, many of them also may be finishing PhDs or their masters. What advice would you give your younger self? It's basically this audience, so that they would start companies or what would you tell yourself?

Chris 22:35 Yeah. So 'cause I, I... So I did a PhD before, um, founding a company and, um, so I was in research and, uh, I found the actual technical stuff extremely interesting, um, just from a technical perspective. But I think by the end of the PhD, I got slightly disenfranchised because I realized that the vast majority of academic research is, um, not actually useful. It's mostly about kind of getting the next sort of NeurIPS paper or something, beating some benchmark by nought point one percent by some interesting method, but it's like, it's not actually super important. Um, there are some papers, of course, that actually are super important, but I think the vast majority of researchers, uh, don't even touch, uh, that kind of research. Um, whereas a startup, the, the reason why I got super excited about doing it is you have this sort of fundamental, uh, gravitational pull, which is that you have to actually get revenue in order to continue existing. So you have to build something that people are actually willing to pay for, and you have to continue growing, and you have to keep iterating. And so you've got this forcing function to actually do something, um, something useful with your, with your time. And, um, I found the product side of things to be the most fun part of doing a startup. And so for me, that was a much better fit. And so, yeah, I would tell my younger self, uh, "Go all in and, uh, you know, don't do a delayed life plan. Just, uh, just get straight into it."

Speaker 23:46 Yeah, I would agree. I think it's very, very easy to slip into the delayed life plan, um, especially if you're at a kind of quant firm, you're getting paid good salary. But if it's your dream startup company, then now is the time. So, yeah, just, just do it.

Sanj 24:00 Yeah. So I definitely considered being an academic, having finished my PhD, and during COVID was, like, applying for jobs. I remember applying Meta, like, so many jobs. Then I remember I got an offer from a startup. Um, they had, like, ten people or so, and the next day I got an offer, and I was like, "Wow, I've been in five interviews," and, like, all these Byzantine processes and the startup has just offered me, like, the next day. And so I was pretty amazed by that, and I thought, "Why not? Let's give it a shot." And so I guess I stumbled into, like, joining as an employee at a startup. That went really well. We were acquired. Uh, and then quit my job to start Solve. So I guess there's also some value in joining early-stage startups as well. So come join us.

Diana Hu 24:42 We'll open it up to a couple of questions. I think we have people running mics. So I'm gonna point, and then we'll go to the mic. So, uh, right there. Over there.

Speaker 24:52 Hi there. Uh, you said you haven't done any sales. Uh, how did you get to the point where you were, uh, getting it in your users' heads that you go see them and go find out what their problems were and what you needed to build at the start?

Chris 25:03 Yeah. So firstly, most of sales is actually not-- Like, I think people have this classical view of a salesperson as being a sort of used car salesman, where they're kind of making a pitch about something. And, um, there are certain times in the sales process where that is the case. But actually, a lot of sales is sort of asking questions and trying to understand... It's, like, exactly the same as what you guys should all be doing if you are building side projects and startups and so on, is asking your users questions about what does their current daily workflow look like, where do they spend their time, what are their actual real problems? Taking that back, making sure that you can then build something, then show the thing you've built in a clear way that they can understand, and that's really what the sales process is about. And once you kind of get an idea about the sort of pattern, your so-called ICP, ideal customer profile, the people that look like the sorts of people that you've sold to before, and you can start doing that, and then you can start only reaching out to those people, get on a call. Suddenly, you realize there's the same sort of sixty questions that they will always ask, and then you're kind of doing the same thing, and it becomes a bit more sort of repeatable and mechanical. But really it's about taking a sort of curious mindset and trying to solve their, their problems. That's fundamentally what it's about.

Speaker 26:07 Hi. Uh, thanks a lot for sharing your experience. Uh, so my question is also a bit sales related. If it's possible, could you share how you signed your first paid customer and what exactly helped?

Chris 26:20 Yeah. So, um, for us, it seems-- So we build AI for, uh, for patent workflows, and there's sort of two main groups involved in that. There's the inventor who comes up with the idea and wants to get the patent, and then there's the law firm, um, who helps the actual-- with the legal process. And when we started the company, um, something like eighty-six percent of the interest that we got, so either people filling in our contact form or me reaching out to people, eighty-six percent of the interest was from the inventor side of things. And so the first person that we sold to was, um, someone that had sort of seen our YC launch during that week. We'd, like, posted on LinkedIn, and, uh, just some random inventor had seen it. And, uh, you know, he contacted me, and I said, um, "Hey, yeah, let's jump on a call." So I sort of demoed this sort of dev environment we had completely local. We didn't actually have it available. And, um, he was basically this solo inventor. He had this idea, and, um, he didn't want to pay for a lawyer. And so we said, "Okay, great. We've got this software that can help you." I really didn't know what I was talking about. I didn't know anything about patents. We kind of built this hacky thing. Um, but I said, "Okay, if you fill out-- If you go to the Stripe link, you can pay us ninety-nine dollars, and you'll get access to this, this platform." And he said-

Diana Hu 27:27 Now you probably charge way more than ninety-nine.

Chris 27:29 Yeah, we charge a bit more than that. This is, like, when we first launched the thing. Um, so he went in, he signed up, and I said, "Okay, great. Well, it's end of Fri- it's end of Friday now, so how about we set up the onboarding on the Monday?" But that was because we didn't actually have a product that we could have him use. So then, you know, we spent the weekend working and, um, on Monday, we onboarded him. He started using it, went off, and by the end of the week, he sent me this angry email because he drafted this patent application that looked absolutely nothing like a patent application. This was sort of GPT-3 or something and, like, you know, with barely any knowledge from our side on patents. And so, uh, we, we then gave him this-- a refund at the end of the week. But what we learned from that is that-

Speaker 28:05 [laughs]

Chris 28:06 -we shouldn't sell to inventors. We should sell to the lawyers.

Diana Hu 28:09 Ah.

Chris 28:10 Because the following week, we then made a sale to a lawyer that I basically got in touch with by me outbounding on LinkedIn. You know, just basically profiling a thousand patent attorneys, send them connection requests, then maybe fifty of them accept, and then you send them a note and say, "Hey, building this thing," or whatever, like a short, concise thing, get them interested in, in you. And then maybe five of them agree to jump on a call, and then one of them pays. And that was the real actual lawyer. And this guy, even though our product sucked, there were certain specific things where if you were an actual lawyer and knew-- who knew what they were doing, you could get value out of it. And so that guy didn't churn. He could kind of give us more feedback, and we could iterate, and we could then reach out to more lawyers and then, you know, sell to them. So that was kind of how we, we started off.

Speaker 28:53 Thanks a lot. So my question is, uh, what's your biggest setback so far, and, uh, what's your takeaway from that?

Chris 29:02 Biggest set-- I mean, there's been a lot of challenges, right? I mean, there's basically every single part of running the company, I would say, is, um, is always, like, difficult, but it's, like, an interesting... Isn't it? 'Cause as, as the company reaches different stages, there's new challenges that you encounter where you have no idea. You've never-- There's no pattern recognition on our part. We've never done this before. I'd say probably the fundamental, um, mistake-- If we were sort of thinking like, "What sh- what do we wish we would have done differently?" I think for us it was we did a very good job, I think, of being lean. You know, when we raised our Series B, we'd raised fifty-five million dollars. We were profitable. We had tens of millions of, of revenue. Um, but I think we should have doubled down on scaling the team and bringing on really talented people much, much earlier. And I think if I was to boil down what the most important thing is about a startup, it's, A, obviously the co-founders, 'cause that's where you're sort of seeding all the talent. But then, B, it's the subsequent people that you're bringing in and how you sort of build out that talent. And I, I think we would've b- we would've been able to scale even faster than we have done keep, you know, keeping bringing the products and so on, if we, if we doubled down on that a little bit earlier.

Speaker 30:05 Thank you. Um, my question was, if you had, particularly for Sanj, if you were at a startup as a sort of founding engineer or whatever, my question for you is, what would be your advice to someone who's working at a startup but also wants to found themselves? What is sort of the top three things you would say is to focus on while you're there?

Sanj 30:29 I just thought it was incredibly fun, the chaos, and then I realized that I really enjoyed it. I think the week I joined my startup, they lost a few hundred thousand from a bad trade. It was like a trading startup, and they only were managing, like, one point five million. And then CEO resigned a month later, and I saw that kind of as, like, an opportunity to be like, "Oh, wow, we can actually, like... I can ship a load of code. I can get a lot of things done. I can learn a lot." And they actually had some interesting customers as well, so I could see every part of the business. You don't get that opportunity as much, and we-- this one thing we try and preserve in Solve is, like, every member of the team can see our revenue. They can see how every part of the business. 'Cause I think that's the difference, right? Like, if you're in a big, big tech company, finance, whatever, you don't get to see the full business, how it operates, how it works, speak to customers, and I think that's the magic that a startup has and why I got, I got the bug, I guess, and didn't wanna go back, um, to, to anything else.

Speaker 31:23 So number two.

Sanj 31:25 Make, make every part of the business your personal [laughs] thing, and then only when, like, founders are like, "Ho-hold on, like, maybe you shouldn't do this," like, really, like, just be like, "Why aren't we closing this customer?" Like, maybe the product's not good enough. Like, just take ridiculous ownership over what you're doing.

Diana Hu 31:42 Very cool. Let's give it a round applause to Chris, Angus, and Sanj. [audience applauding] Thank you so much.