321. The Exited Founder Building AI for Due Diligence With 7-Figure Revenue w/ Yi Luo | Eunice

0:00So my name is Yi. I'm the CEO and co-founder of Yunus. Yunus is really around the information asymmetry on the private market side. So essentially we enable the AI to help our end users to essentially do diligence and monitoring the asset level, very much specifically focus on the private market.

0:24Amardeep Parmar:And then you recently raised around as well by Moonfire, right? So where are you at the moment? Like what's the progress? So we raised around end of the last year and it's been a really interesting time, I think, as a founder to raise in the VC market. I do think nowadays attraction is more important than anything else. Obviously we keep growing with the fund that we have raised. Now we move from our core market, which is a digital asset, into a much broader private market now. And you mentioned that obviously because you're a second-time founder and you've already built a company before. And usually what people say is that's going to be a huge advantage.

1:00Amardeep Parmar:So what was your experience there when people can say like, okay, you've already built another company beforehand, you've had the exit. And was that not as big advantage as you hoped you might be? I really feel like it's getting less, I think, advantage five to 10 years ago. So I think the argument 10 years ago is whether you made the mistakes in your first time, I mean, you learn from a mistake, you try and make less mistakes in your second time and all this stuff. I do think that skill applies, never black and white, but because of how the world of operating and growing a business have changed so much.

1:42So as a founder, as a CEO myself, and also my CEO is also a second-time founder, a lot time we've been thinking about is that what is the learning we have from the past startup that we need to throw away? And it's basically the legacy that is not useful anymore in this new world of operating. And what is useful? So we think about it every day. And I don't think that blankly just becoming a second time founder is going to be necessarily an advantage. I do think about another theme which a VC invested into the founder who has childhood trauma. I think that thesis makes more sense because it makes you a lot more resilient.

2:28The second time founder has advantage, but I think they really need to put away the things who are not useful in the nowadays.

2:36Amardeep Parmar:And you're saying that you had to show traction to get this around, right? What kind of traction did you build and what did you have to show to try and get invested in it all? Yeah, so number-wise in the last two years we've been quite stable around 2.53x year by year and we are currently on the seventh figure in terms of a revenue. As you could imagine we've done this zero to one journey and now very much focus on how we bring this into a scale. And you're saying how you have to throw some way some of the things you did in your previous company, the way you're thinking and the way you went about things.

3:15Amardeep Parmar:What are some of those things that have changed and you now can't think the same way anymore? I really feel like the biggest change as a founder, I know that people who listen to the broadcast, they have experienced themselves, it's a notion of like, you could really lean on a very lean team to build a big operation. I think this is really fundamentally change a lot of the founders how we operate. and thinking about this AI layer within your operation. And also, there's a lot to consider based on your particular startup. Because in our space, we talk a lot with regulators, actually. We also talk a lot with B2B, that large enterprise clients.

3:57There's only so much you can automate at the same time, and you will have to apply human layer on every single thing we do. And this is actually like a lot of our clients paying that for this reason. So I think to reconsider that what applies and what not applies when the Claude and OpenAI and all the general model becoming super powerful is the biggest priority for all the CEO and founder.

4:28Amardeep Parmar:And you said as well, like some of your clients are massive, right? And how did you win those clients at the beginning? Like how did you get that first big client? I really feel it is one thing for us to get big clients in the door is to focus on being the top expert in the niche. So in the unit, I generally operate in the system that I feel only two types of AI startup going to thrive. One is a general model by the Anthropic Open AI, and there's going to be more general models appear. and they progress really, really fast. And they are the assistant, they help the human to operate and become a more efficient expert.

5:15How we build in the units, and this is why I generally believe if another type of AI startup is going to thrive, if you are actually the expert in what you are doing, you really can guide your client in the entire process end to end and also help them on the stakeholder management, such as bridging the regulator and the industry on a certain practice. So in the UNIS, my playbook, no matter if it started in a digital asset and then later move on the private market, it very much focused on being the expert, hire the expert to begin with and then grow from there.

5:55Amardeep Parmar:And before this starts with Zara, so you're investing for a while and you said how you're trying to think about what's the right path for you now, right? What was it about Eunice and the problem you're solving here that made you decide this is out when you become a founder again? Yeah, and go through all of the pain that comes up becoming a founder, right? Yeah I think this has never been you know accident Because I've been always operates in a very gray market I started my career in a quite unconventional path. So I started in the art world used to being auction house and the gallery, which is quite rare nowadays, looking at tech world.

6:33And then after that, I moved into a VC for some time. And after a VC time, essentially I built my first startup. I also do Android investments, so I have a lens of also investing into some funds, some of the startup. And one thing I've been doing, I think the first 15 years of my life, is to hunt the information that somebody already has. So really who holds the information as the one that I would say more empowered, a lot more empowered and also likely going to win in the end. And eventually that really strikes me. It's not really me or anyone that hangs in the information in the market. It is how the private market is being designed.

7:20It's a system. So right now, the problem I'm fixing is to build a better intelligent layer using the power of the AI. I think unlike some of the last generation, how the data, how the intelligence layer has been built by the human hands for decades, leveraging AI, you can do that a lot faster and in a way that really empowering people and also empowering these assets to reach much broader capital and access.

7:52Amardeep Parmar:And this is Miranda's why you've got a co-founder who's also known as company beforehand. How did you join forces and what made you decide to work with him? Yeah, so Philip is my CTO and he also a great founder himself. He previously built an AI computer vision startup from San Francisco and raised$40 million, a Series B still alive. And we essentially come, like met each other actually from the Asia community as well. And we were comparing the nodes. We have essentially built a personal relationship and it come all very natural at the point that I need a tech leadership within the team. And on his side, he previously worked for a big scale up as a head of engineering, which is a good node.

8:42I don't know if you heard about them, they have like 30 million users. And then they've just come quite naturally that we are committed to fix this problem together. And I think, yeah, I think choosing co-founders is such an important part of life, which I could go on for hours. It's much more difficult than choosing your life partner even. But I think for now it's been a great journey.

9:09Amardeep Parmar:And then working together, what do you think makes our relationship stay strong? How do you manage that? We both are very intense founders. We both are opinionated. We really care. I think the good thing is that we care from different angles and we are very complementary from what we do on our skill set. I think fundamentally, this is all about, I think I read a statistic somewhere, like 60 % of the startups fell because of the co-founders relationship fell. And I think the best co-founder are those ones who you know you trust, unconditional, on a personal level. He's still your best friend if you break up on a business decision.

9:49You know that he actually care you in person. I think those are very, very long lasting because I have argument with my co-founder all the time on the business decisions, on the clients. We can always come back in a position that I trust you have the best intention and it is more of the opinion piece we need to reconcilate. I do have worked with my CTO in my previous FinTech as well. We have gone into acquisition the entire journey and he stayed with me in the first part of lunettes until he decided to step away and then essentially building his own gaming. And I think the relationship is very much similar there.

10:39To trust each other on a personal basis and really, it's not really a business relationship when you're a co-founder. It's very much going through the lowest point where fundraising is becoming really tough. do you look at your co-founder and smile and bring him out for a lovely dinner in San Francisco. I mean, you can still really enjoy the time together as a human being. And that is the one you've got to look for. And I feel like I'm probably lucky that I have my co-founder on that basis.

11:19Amardeep Parmar:Awesome. And then how big is the team now? Only 12 of us. Yeah. But you've done that in quite a short amount of time as well, right? How did you pick those people? Because obviously they got to compliment you and your co-founder too, right? Yeah. What do you look for? I think it's not one size fits all. Obviously, my co-founder, who was sort of in charge of a tech hire, I'm a lot more on the commercial and business side. I generally feel like the future of the talent to some extent, it may be a little bit extreme, but I think you either have to be a builder or you have to be a seller. So, I mean, ideally both.

11:56I do think now the bar has raised massively. The engineers now we have seen who also done an excellent job in selling and some of the people who work with the commercial side and also can vibe code and bring a prototype into a very different shape. I do think that that type of talent and the talent who has ownership, I mean, I know everybody is looking for talent and have ownership. But when you find one, you can truly feel it. It's where we are hiring. I definitely think the bar of hiring has been rising massively in the last half months.

12:34Amardeep Parmar:And there's been a debate recently as well, right? Where I think it was Uber last week, you were saying about all of the money that's been spending on like AI tokens, like how that hasn't necessarily paid off compared to like hiring more people. And then obviously you're having to balance that out as well, right? Of like how many people do you hire? What do you use AI for? and how are you using AI internally right now? Obviously, we use a lot of AI. We do... I mean, I think one thing quite important is that the final result we give to our clients, like in many cases, it's either that being very thoroughly checked internally or it's very, very highlighted that clients may have to check a certain section.

13:19So, we put a lot of guardrails in the product we've been building. And we also use AI ourselves. I do think that with the speed of the model being progressed, you have seen there's been amazing results come from AI. It's like, okay, you know, two years ago, this will take my head of research for entire months to build and now can get it done pretty much in one night. But there's also a lot of the shortfalls of AI. I think I heard about the same thing in the coding side. Startups just completely rely on the coding. But if you actually read what AI produced line by line, even if you give it the entire context from your intelligent layer, there are still, I would say, a decent amount of things that AI can miss.

14:11And I think I really feel like the future of the world, the people would buy a validation layer. So that is actually where becoming the most value comes from, because generative is becoming too easy.

14:25Amardeep Parmar:And then when you're kind of now selling, right, you said like you've got the sellers and you've got the builders. And I'm guessing you're more on the seller side, right? And like selling is like something which is so hard for so many founders, right? So when you're going in to frame it to these companies, what framing did you have to? How did you work on that storytelling to make sure that you're getting these people to give you a try, especially at the beginning? I can assure you, in the beginning, I think I'm the worst self I could have, in my entire people I know. Because in my last FinTech, I'm much more of a CEO role that I hide behind the CEO with amazing self.

15:07and I would never think I could do B2B sales when I started to be a CEO of Eunice. But over time, quite interestingly, I do find I develop my own style. I'm really not a person who could do a hard sell. I just try generally being helpful. It's not an easy journey, especially somebody who is probably coming from a different culture, come from a different language, and then been dropping to the enterprise B2B South code, which I've never done before. But over time, I think people, like enterprise South is a people business. It's really about focusing on the problem. What exactly is the problem that with your niche, so perform initial five users, they're trying to really fix and really being caring and being thoughtful about their problem.

16:00I think I've been very lucky that I have encountered a lot of problems in my previous journey that I've used to do due diligence and monitoring when I was a VC, when I was doing the investment. So a lot of the problems to me is not really unfamiliar. I think that is basically my style of the sales. Sometimes I still watch people who actually come from a sales background and I feel like, wow, the delivery was perfect. I hope I could do that. But then I think over time, and I also had I was chatting with some of my engineers, is that actually we all have our own self style. It's more like how do you develop and trust your own style.

16:45Amardeep Parmar:Hello, hello. I hope you've enjoyed the show so far. I'm Amarit Parma, co-founder of BayHQ and the host of this show. For those of you who don't know, BayHQ is the community for high growth Asian heritage founders and investors in the UK. Over 7 ,500 people have attended our events. 200 people have been for our impact programs. And obviously, there's been over 250 episodes of this podcast. If you want to join us and take part in the programs and come to the events, go to bayhq.com forward slash join. We also just released our first ever book called Startups for Outsiders, which you can get on Amazon now.

17:24Amardeep Parmar:And the link is in the bio. Hope to see you soon at our event. Hope you enjoy the rest of the episode. One of the interesting things too now is how pricing is changing so much. And about, I think a lot of people get mixed messages, right? When they're trying to get their first customers and somebody's telling them do price based or usage based or do seat based. And how have you kind of navigated that and what's the right model for your future? I think we really aim anchor on the value based. I think this is where also as an AI startup, you will have to very much focus on the result and the value you provide and how much clients would pay if they go alternative, if they build their own, if they have an internal team handle it, if they go to a law firm for it, etc.

18:11So we've been very, I think, lucky to work with some enterprise that really have been spent a lot of money, a lot of efforts on a particular process, and we've been able to close them on the relatively large contract. I think it also proves the value we provide to them. It's not just the AI layer, but we also help them on essentially from all the way to maintain the template, maintain the change in the space it operates, and connect the due diligence with monitoring, the maintenance on the data side, and essentially also bridge the stakeholder they have, which is other regulators. So I think we basically cover the entire value chain on a very niche problem, and that is what the clients are paying for.

19:00Amardeep Parmar:And we talked as well about LPs and VCs earlier, right? And obviously, what we're doing and like, potentially, hypothetically, the fund later on in this year, then it's like, we've got to be careful what we say, right? How does it work for an LP, right? If an LP is using Unis, how would they interact with it? What would it do for them? Yeah, so we essentially have two platforms. One platform was built for digital assets, and that was our brand, Butter, to begin with. Now we're expanding into a private market. So we have a separate platform that purposely built for LP workflow. And it started from the due diligence and also the monitoring.

19:37So the module and the logic there are similar. And we very, very much focus a lot around the validation. So this information you get from the GP, from the market, that what the gut rails, what the data you have, and also your process to help you to validate the claim you have. is what we've been put a lot of thoughts when we build the product itself. And also we have a human layer to actually help our client to validate on top of it. And also when it comes to a portfolio monitoring, there's a lot of a problem around the documentation, the date, the decision you have made pre-investment are very disconnected after investment.

20:25and then to understand the change of it and also to get a view around among all the funds, all the investment that I've invested, that was, for example, my healthcare exposure, what happened if Claude got lifted, that impact three funds I've invested positively. And all these questions now could still take a human like two weeks to figure it out. among the sea of information. And I think also, interestingly enough, that I think asset allocators are both ones who hold actually a lot of responsibility and power in the industry. But I do really feel and see that they're still very underserved by AI and also responsible for AI.

21:16So the AI that designed for their purpose to be able to see through the private market, to get as many information as possible, to get as many validation as possible, and then to bet the best manager in the next generation to really look into the best asset. So I think a lot of the LP we're taught to have a very lean team, as you're probably also aware of. I have a small team, but they manage a massive amount of assets. So that's how I choose to go for asset allocator on this expansion as a first step. because I just really generally believe that we're underserved.

21:55Amardeep Parmar:It seems to be said that as well, because when it's such a huge amount of money that's being managed by a few people, do you think it's because of the people relationship of where somebody only trusts one person to look at it? So they wouldn't want to hire more people, but they will pay more because they've got the money, right? If you're like, if you're in charge of like a£10 billion fund, you could hire more people, but they choose not to for some reason. And it's kind of interesting, like why do they not have the bigger team sometimes? From our observation, from a third party point of view, is that a lot of LPs also have a background from a public sort of angle.

22:31So there could be a pension, there could be an endowment fund. It's quite important for them to stay relatively linked and at least very thoughtful on the cost of structure. So we do definitely have some LPs who employ a big team. But even that, they are also looking a lot around automation, how to keep the team lean while they can expose to more opportunity. So what we are also building is exactly to help them and to empower them to be able to look into more opportunity, spend more time talking to a GP, while the things that are more administrative, our data cleaning, our validation side, that you have a purposely built AI to help you to do that part.

23:19Amardeep Parmar:It's also good, I think, as well, from the industry's perspective, is that if the LPs, at the moment, they do things in a very old school way, like stereotypically, right? And if they're doing things in spreadsheets, but then they're then investing into the fund managers who are using AI and doing everything on their frontier edge, that disconnect is probably, it's going to more like cause misallocation of resources, right? Whereas when they're also using frontier AI resources and they're also looking at AI in the same way, it can then mean there's less of a gap between the LPs and GPs and how they're operating?

23:49Absolutely. And I think a lot of LPs are very, very keen to move to a world that they can ask good questions much faster in a way that really meets their fiduciary duty as well. And I also think that thinking about LPs problems, they invest into many founders, like hundreds of thousands of assets underneath of it. A sea of information as LPs got to process. It's all over the next level. So I think to many extent, they are actually the perfect use case for AI, whether if it's that purposely built, responsibly built, help you to filter out the noise, your risk-based approach and understanding that what is actually moving the needle.

24:31And also, very importantly, pass that learning for your future investment, connect that dots, really through the data layer. I think it's very, very important for them.

24:41Amardeep Parmar:And you said that this is a separate platform, right? It's completely separate. One of the challenges for any startup is when to do another product and when is the right moment. And how did you decide now is the right moment to have a separate platform and run two platforms at once? I think now to building a product, to building a platform, honestly, the cost has come down a lot cheaper. We obviously have all the SOC2 and security measures you can think of in the place. But I really feel those two types of clients we have are very, very different. There are a lot of product logic are similar, like the validation and then how do you put the guardrails on the data, on the citation, on the sources.

25:23But I still think the workflow and how they deal with the nitty gritty are very different. That it's just like not make sense to jam them into one platform, it much, much makes more sense to build for their own purpose.