E08: Ilya Kravtsov - Serial Entrepreneurship & AI Banking

Ilya Kravtsov: If you don't move fast, you'll be out, right? The company is going to die. Just say very simple, if the company doesn't grow and it's a tech startup, it's going to die. Right? The outcome is quite binary. Projecting a hundred million in five years business plan like everybody used to do before is not good enough. If you are in global AI space, people go to 0 to 100 million IRR in like 10, 15 months. Right? Or even less. I think the branch network is going to disappear. Right? That's already happening with digital banks. But even more, I think all the apps that the banks are developing are also going to disappear. Banking will not be customer driven, but will be AI driven where you are recommended a product at the right time and the only thing you need to do is say yes.

Felix: Hi everybody and welcome to a new episode of True Founder Stories. This is a show where we talk about founders, personal backgrounds and discuss FinTech AI and cross border expansion. With me today is Ilya Krafttsoff who is one of the co founders of Rinkus. As you mentioned, uh, our show is about fintech AI and cross border expansion. This man is doing it all. Ilya, thanks so much for being on the show today.

Ilya Kravtsov: Thanks Felix. Good to have you.

Felix: Let's start learning a little bit about Rinkus. What do you guys actually do?

Ilya Kravtsov: So Rinkas we started roughly about four years ago. Uh, and uh, the premise was, uh, you know, we looked at different uh, sectors within the fintech space. We found that there was a big gap uh, around home ownership in Southeast Asia and particularly Indonesia where um, one of the key indicators was the mortgage to GDP ratio, which is very low, um, in particular for Indonesia is below 3%.

Felix: Right.

Ilya Kravtsov: And then we started like we thought that that was a bit really an odd indicator if we compare to more developed markets like us and so on where it's like 50% and above. Uh, so we started really digging deeper understanding, you know, why this problem persists, uh, can we do to facilitate that. Uh, that's how we kind of kicked off everything. Uh, and we went through a lot of exercise of really talking to the large players from both on the um, bank side and property developer side to understand what was the issue and how could we unlock and increase that indicator. Right. So that's how Rinkus more or less started around four years ago.

Felix: Obviously you've secured over $10 million in funding at a time where in Indonesia very few other players were actually able to raise. Um, what's the new direction that you're taking?

Ilya Kravtsov: Rodriguenka is now I think uh, it's a natural evolution. So I think the way I look at startups is you always have that, you always have like a, it's like a steps approach. Right. So you start a business, you have certain assumptions, you need to prove those assumptions or disprove those assumptions and then you, you know, if everything goes well you experience like a good growth. And that's what we, we went really when we started the business, uh, we you know partnered with large property developers. So that was actually our growth hack where we secured large volume right of inventory and then that unlocked us all the relationship with the uh, with the bank partners. Right. Because as a startup is very hard to go and convince the banks to work with you.

Felix: Right, of course.

Ilya Kravtsov: Um, but by having very large uh, supply we managed to get those relationship and literally when we kicked off the business we went from zero to like in 18 months from, for more than a billion dollars of transactions uh through our system. Uh so basically mortgages dispersed and that was really intense and great period. And then we've been uh, continuously thinking okay, what is the next kicker of growth, how we can grow this father. And of course as a founder there's different avenues that you can take. Um, for us we always uh, we're quite an international team so we always thought that the business will be more than just one country. Right. But we wanted to make sure that in that one country it kind of works. We become the market leader. It's like we already know the ins and outs and the product is also mature. So that's where we got into the stage in Indonesia. Indonesia we are by far the number one in this space. So we're the largest mortgage infrastructure platform in the country. Uh, if we were a bank we would be probably in the top five banks in terms of volume of mortgages that's going through us. So I think that is, that is something that we are proud of and it was a lot of work but of course you know we're always hungry for more growth. Right. So I think the avenues that we decided uh, that gonna unlock us more growth is one is the international expansion. So we already generating revenue from uh, Middle east at the moment. Uh and uh, the second one is also during this past four years the whole world went through a uh, technology revolution and is still going through. This is something that we cannot ignore and this is the second uh, avenue for growth for us. It's um, going more towards the AI approach.

Felix: And you previously actually said that found often go where funding flows do you find yourself going in that direction with Rinka's itself, or is it sort of a natural progression for you?

Ilya Kravtsov: I would say it's not as straightforward in that sense, but, um, of course you want to be, I mean, let me backtrack a little bit. Right. So you um, want to be in general and regardless startups, you want to be in a place where there's a lot of excitement, there's a lot of optimism, there's a lot of things happening. Right. Regardless of what, what do you do? Right. Because of your job, actually. Right. Um, so in that sense, if you're a startup, you want to be in a place where there's a lot of new technological innovation, there's a lot of investment happening around that, there's a lot of excitement, there's a lot of push from the government. Right. So you always seek that kind of ecosystem spillovers, um, that can help you. Right. And that is naturally that anybody would look for. Right. So we also look for that. And we experienced that early on in the days in Indonesia where it was really, uh, booming and um, growing very fast. And we see some of this in other parts of the world now in the US of course, in the AI space is growing really very fast, I would say in the Middle east, before uh, the war, and hopefully it's going to be sold soon. It was also really growing from the investment and uh, perspective and AI investments as well. So that's kind of like a natural, I think evolution that anybody in any business would look for.

Felix: Yeah, absolutely. And I think like in particular, vertical AI is really top of mind for many investors. Uh, I think there's two major big themes. Right. Vertical AI continues to be strong. The other is sort of AI native services. And a lot of the large investors have been talking about this. I think the application layer is particularly interesting on the foundational models. I think our thesis from the very start was that maybe we sort of exaggerated a bit by saying it will commoditize. At least what we're seeing. It's very competitive. Anthropic is overtaking OpenAI and revenues in, uh, a very short time frame. Things are very volatile. But on the application layer you have very sticky customers that just need solutions. The penetration for enterprises and, and vertical companies is, is very low actually when we look at AI. So um, we're very excited. There's very different models. Right. There's AI for hotel operations, voice agents for these guys. There is AI for dentists. Right. And these are excellent customers to sell to because they understand the value Proposition very quickly and they have pretty high retention and low churn. Right. These are some of the factors that we're looking. What are the opportunities that you see in fintech for vertical AI?

Ilya Kravtsov: We've thought of course about it internally a lot and I think I have also a personal view uh, on um, the battle between LLMs and the so called apps layer and so on. And I think just to give you a bit of a historical background, so when I started my first job, uh, iPhone one just came out, right? So I'm not that old but when people think about iPhone one is like somebody is like history, right? Well that's where I entered the job market and at that time uh, the world was experiencing a revolution which is maybe not technology revolution, which is maybe not as big as AI, but it was pretty big at that time. And that revolution was around mobile phones, smartphones about um, being able to um, download apps, communicate and having all these telcos rolling out the network so you can basically send messages from anywhere and so on and so forth. Those were the days. And I always compare the current LLM foundational model companies and apps layer to a little bit what happened 20 years ago with the telcos. So telcos uh, were kind of similar to the foundational models in certain degree, right? So they were the ones who would um, have very high capex because they need to roll out network, right? And at the beginning uh, the network was not yet commoditized. So if you're like with one operator you have coverage maybe everywhere and the other operator you only have coverage in the big cities. So there was a big difference in how networks, uh, the quality of those networks back in the days and I think all the large telecom operators have aspiration, had aspirations at that time to enter the apps layer, right? So similar kind of what is happening now with the foundational models, right? They're all very capex intensive. The models are still fairly different, evolving uh, fast and all of them also have the aspirations and to go into the apps layer. Now if we compare what happened, you know the good thing about history that you know what the outcome was, right? So the outcome in case of telcos was that a large telecommunication uh, companies uh, failed in going into the apps layer, right? So they didn't manage to launch any significant services of over the top, uh, they eventually gave up on it and eventually focused on their core business which is network, uh, network got commoditized, right? So everybody, there was no major difference now between using one or another, right? And then the price war started. Right. And you know, we could argue why it happened, why they failed. I think there's a lot of reasons. Maybe some of them linked to organizational struct, some of them linked to many other things. Right. So the big question mark that I have now, what is going to happen with the battle between the LLM foundational models and the apps layer today? Right. Is it going to follow the same path I think followed with the telcos? Is it going to be different? I think, um, uh, it's still, we don't know the outcome. I mean I would say last year I was convinced it would be following very similar the trend of telcos. This year I think Claude surprised everybody how good they are in product for sure. Um, but having said that, I think the fundamental belief on vertical AI is still the fact that all the vertical AI spaces are still open, except I would say probably the coding one. Right. The coding one is already very crowded. You have cursor, have cloud call, you have codex. And this is where probably 90% of the application revenues are coming from at the moment. You can see actually from all the other verticals it's still very early days. Um, so fundamental belief here is that on the financial vertical is a difficult one to conquer because uh, there's a compliance aspect, there's a risk aspect. You're working with large banks. So for players, LLM foundational models, it will be hard to adapt that in every single legislation country and you know, it's just a lot of customization work and it's still a very open segment. Right. So our fundamental belief is that focusing on vertical AI in financial sector early on will give you that stickiness and that advantage that eventually Even the large LLMs will struggle to fully replace. Right. Uh, so that's how, where we are starting and I guess that we started by working with the banks from a mortgage angle. We have now 33 banks and I think now we're going more and more into that uh, vertical AI aspect.

Felix: Yeah, and it's pretty interesting because Sam Altman also provided that analogy where he says he's going to be like utility provider, like electricity or water and people will use those large providers and those vendors, um, like they now contract electricity or anything else. And that would be the same for intelligence. But um, I think it's worth talking about the different technologies as well that will kind of drive this whole change because I think the first thing that we saw is 2022 when conversational AI really came um, to light. 2024, arguably with Whisper, um, and 11 Labs et cetera, voice AI became really dominant. And that's also another short term inflection point which enabled a lot of these business models, uh, as I mentioned, um, for dental clinics, receptionists, for hotels, concierge and all these different things. Uh, and then 2025 is really where computer use agents kicked in and suddenly you can use classic uh, and traditional ERPs and connect them to modern systems because now the computer use agent can basically cross that bridge. That used to be just impossible with APIs because APIs kind of sucked and weren't really built out and to a large extent still aren't for many verticals. What are the specific technological elements that uh, you're adopting for rinkize right now and what do you think is most exciting in terms of shift?

Ilya Kravtsov: There's a few things, right? One is the trend that we see now for what we call the apps layer companies. Right. So the trend of the apps layer company I would say maybe last year and it's evolving extremely fast. Right. Last year would really focus on customer acquisition, focus on UI ux, uh, focus on all that right now. I think the trend that we see this year of apps layer companies is um, and we can take some examples like Lovable or you know the hyper, hyper growth companies. I think what they realize and also similarly happened to Cursor is that they successfully managed to capture a large amount of customers but they're still relying on you know, the anthropics um, of the world. Right. For their LLMs. Right. And that is becoming more and more of a friction situation. And we've seen that happened with Cursor and cloud code. Right? That's a very, very simple example. And this is just in one vertical, right. So now I think Lovable is the next vertical where we already see artifact inside cloud uh, workspaces is really getting there. Right. So the question is uh, so the trend that we see now is that a lot of those, I would say app layer companies don't want fully rel on the anthropics of this world but they're actually looking at taking the open source Chinese models and doing additional RL on top and making their own. Right. So I think that is like the high level trend we see. Right. Uh, from our perspective I think uh, of course doing RL is kind of expensive, right? So I would say we are not yet as far in that spectrum. Right. But where we are at the moment is we have an uh, AI orchestration level that definitely uh, so we are at iteration number three. So first one was Risa 0 then we have Risa 2.0 and Risa next is coming. I think our internal evolution within the company is we started with very simple one model, OpenAI based, uh, our internal AI. And we are now understanding that one model, of course it's not sustainable from both cost perspective but also other models have other functionalities that are better. So we're launching a multimodal platform where we can use and can orchestrate depending we want. This is more on the LLM side what we're also focusing. We understand that different clients will have different needs. Right. Ah, so we call. So it's a multi tenant type of approach. Right. And you want to make sure that these needs are catered very quickly on the fly. Right. So if you want to launch uh, let's say financial um, AI for a certain type of high value customers, you want this AI to talk in a certain way because they're high value customers, they want to talk about certain topics, they want to have a certain way that you interact with them. So we want to enable that. And then most uh, important thing, you also want to make sure that your AI gets always better. So the eval techniques are extremely important and that's what we are really focusing on. So that with this multi model, multi tenant, which becomes a lot of different variations, you can always assess them and the model can auto improve yourself. Right. So that's the core focus now. So as you can see from very simple, kind of like one model, one AI approach, you're getting two multimodal, multi tenant and multi eval approach. Right. So it becomes a lot more complex and that you need to adapt to the realities of large banks where the compliance and so on kicks in.

Felix: Definitely. And maybe help uh, our listeners understand what are the concrete use cases for Rinkus that are live now and what you're planning for the future in terms of interactions.

Ilya Kravtsov: So on the financial sector, as I said we had access, we have access to 33 of the largest banks and we are talking largest banks from Singapore, we work UOB or cbc, we work with icbc, we work with ccb. So largest bank from China, uh, May banks, AIMB from Malaysia, Indonesian banks, BCA Mandira and so on. So we have a good spectrum. Now what we realized is that, and essentially traditionally we've been providing like financial infrastructure mainly for collateral based products for those banks. And we have 1,400 branches actively every day logging in into our platform. Now where we see the evolution is that uh, since so there's three points that I would like to mention, how we got to this point one is traditionally we started with mortgages. A mortgage is one of the most complex financial products. What that means is that usually the way mortgages are sold is through human interaction, right? And that was uh, always the case. Even in the SaaS era. Even if you have SaaS errors, we still, you're buying a $500,000 house, you need to get a mortgage, you want to talk to somebody, right? That's kind of like now I think for the first time in history with AI, this is changing, right? Because where AI is actually good at is in that human aspect. It makes you feel you're actually talking to a human. So one fundamental thesis is we've understood that for the first time, you know, banking can change even with regards to more complex products, right? Because let's say on the more like simple products like personal loans or credit cards is already quite straightforward and can be digitized even using the old SaaS techniques, right? So I think what many founders also make mistake, they put AI everywhere where it's needed or not, right? And that's what we didn't want to do, right? So we wanted to make sure that we are going towards the AI direction where AI has a real differentiator, right? And I think in that, you know, in being able to sell more complex financial products, I think that's where the real differentiator comes in. So that's point number one. Point number two is what we are starting and the first product that is very successful at the moment is we have a lot of bank partners that are sitting on very large amount of customers databases and those databases traditionally are not very active activated, right? So we have one client, has 30 million customers and we always ask them, uh, how do you cross sell or upsell to those customers any products? And they're like well look, we do some data analysis and then we try to segment and then we extract some Excel files this, and then we pass this Excel files to our branch network and then to our call center and then the relationship manager in the branch calls them or the call center calls them and tries to reach out, right? So that's the traditional ways how banks try to cross sell or upsell. Now what we are uh, now successfully doing is said, look, how about instead of using the traditional way of selling, we prove you that AI can do the same with cheaper cost and high impact. So essentially what we do is uh, our multi agent architecture profiles all the customers, segments them beforehand and then our loan recommendation angel already matches them with a product that would be automatically Approved. And remember, those are existing customers, so you already know everything about them. Them. And then what we really focus and AI focuses is on the selling part, right? So not customer service is really selling part. Because eventually if you can add the top line to the banks, that's what they really want, right? And banking is becoming more and more contextual. So you want to give that offer when you really need it, right? And when you offer it, when you really need it, your conversion rates will be going much, much better.

Felix: I think this intersection of proprietary, valuable data and the use of AI, be it chatbots or voice or other means, is where the actual value accrues. Because I see a lot of pages, you go there, there's some sort of AI chatbot. It doesn't really help me because it's basically a glorified lookup.

Ilya Kravtsov: No, agree. And look, if I want to portray, if I want to look a little bit in the future, right. Like before we were talking about digital transformation and the digital bank. I think we need to start talking about AI Bank. And what does AI bank entail, how it's going to look like. And I have a pretty, you know, firm vision on how I think it's, you know, in which direction it's going, how long it's going to take, we don't know. But I think the branch network is going to disappear, right? That's already happening with digital banks. But even more, I think all the apps that the banks are developing are also going to disappear, right? Because the fact that the customer needs to open an app to request a product, fill in a form, it's very customer driven, right. And I think the banking will not be be customer driven, but will be AI driven, where you are recommended a product at the right time and the only thing you need to do is say yes. Right. So that's where I think it's evolving. Of course, we're still far from that, but I think eventually it's going to happen.

Felix: Now banks are spending millions building digital experiences, right? Why do they go with you instead of building stuff themselves?

Ilya Kravtsov: Ah, first of all, um, okay, let's just draw the baseline. So the baseline is that AI is still very new, right? The baseline is that, that nobody's an expert in it, not even the guys in the Silicon Valley. Yet it's evolving so fast that today you might know something, tomorrow might be completely obsolete, Right? So it's growing very fast and nobody knows now what is happening on the banking side. On the banking side, the banks, they have a push from the top to start Using AI. Right. Uh, and that is very clear now to develop their own and to really understand, you need to build certain capabilities which usually actually the bank's IT capabilities are very limited. Banks are not tech companies, let's be honest. Banks are companies that are managing risk and they do it very, very well. But they're not tech businesses. So the first thing is banks are looking for players that they already have trust with, uh, they already used in the past, uh, because for compliance reasons and so on and so forth. So for us, the reason number one is we already have access to these banks and they're really been using us for the last four years. So it's very easy to understand what the issues are and they're easy to have that trust element. The second one is, of course, you need to prove that it works. Right. And with us, with relatively limited effort, we can basically make a business case out of it for them. So you don't need to go shop around, do RFPs and reinvent the wheel. You just have somebody you can call. We come, we try it out for two months. It works. Great, Great. Right. So it's just a quick win with an established and trusted player. That's why they use us at the moment.

Felix: Yeah. I want to talk a little bit more about your personal history as a founder as well, because you're actually an entrepreneur that has started multiple companies. At what point did you feel like a serial entrepreneur? And uh, you know, how did your self perception as a founder change over time?

Ilya Kravtsov: I started my first company. I was uncertainty whether to do an MBA or start a business. Right. Um, but I thought I already had two masters. MBA is probably an overkill. And I was already in consulting, management consulting. So like, majority of the people after MBA go to management consulting. So I'm like, maybe that's not the right way. Right. So I started my, my, my first business just because I wanted to learn. Right. And I said, okay, instead of wasting money and doing an mba, I'll, you know, put the money in and build a first business. Right. So that's how it kicked off, right. Really now at my third business. And of course the, you know, I've been doing it for almost 15 years. So it's, it evolved. Right. So the evolution here is that now, now I see being building companies more of a profession. Um, and what I mean profession is that eventually you reflect on your career and you think, okay, what I'm good at, right? And some people are good at being a doctor, giving cures to people. Some people are good at Ah, doing accounting and I think what I've probably specialized and I think I'm okay in going to that zero to one phase and going to adopting the new trends and new technologies and really discovering, you know, problems and how you can solve them with the tools that are available out there. Right. So that's how I see myself now. Right. It's just, you know, that's kind of the evolution from really the first days, which is like learning to kind of like seeing that this is my profession.

Felix: I also think that in past formats you've said that you're chasing growth and milestones a little bit more. As a first time founder, how has that changed over time? How do you kind of look at yourself now?

Ilya Kravtsov: I think when you start the first time you think you know the concepts but in reality you don't fully grasp the concepts. Right. Things like marketing, market size, things like you think that everything is a big market size, which is totally not true. Right. Um, and then you think that you don't understand what is also, you know, going aggressive. Right. What level of revenue is big, which small. And you know, how like there's all these concepts, it's like you understand but you don't fully, you know, uh, digest it right now when you are a bit more mature, you're as uh, a founder, you understand, you compare yourselves more to what is out there there. Right. And I think it's very important to do that comparison to talk to other founders that are successful because that gives you a perspective that if you don't move fast, you'll be out. Right. The company's going to die. Just say very simple, if the company doesn't grow and it's a tech startup, it's going to die. Right. The outcome is quite binary. Right. So and as I said before, there's always a step by step approach. You experience a growth, great, but you cannot ride on that growth forever. You need to have that second level of growth, that third level of growth, etc, etc, etc and usually many founders that are first, younger founders might fail on that second step or third step. Right. Because they do something, it's working great and then they get stuck. Right. So I think the flexibility, mental flexibility of download this next stage and third stage of growth is extremely important and it's also extremely difficult. Right.

Felix: Is that one of the biggest changes that you saw in yourself, that you are iterating more, pivoting more, testing new models more?

Ilya Kravtsov: I think what uh, we understand is, I would say personally I understand more the macro trends than I was before when I was a younger founder, I would just really go the tunnel vision on the product, on my customer, and I would really ignore all the rest. Now, uh, in the world we live, and if you want to build an AI business, you really cannot ignore what's happening around the world. So you have really a macro perspective, first of all. And I think more and more companies, tech companies, are becoming global companies, less of country specific regional companies. And I think this is a trend that we've seen, you know, we can talk about later in Southeast Asia, right? Where before, you know, one country story would be enough. Uh, and I think now we really live in a global world, so that macro perspective is extremely important.

Felix: In the past, as a founder, you've talked about being maximally truth seeking. As a founder, right? And I think it's interesting because, you know, we've been or been in consulting and I think this is a good example where incentives might be different. I also talked about this with Nate, uh, who was on our show before. He, um, talked about this. As a consultant, you provide value, but then there's also corporate dynamics and all these things happening. So there's incentive to not always be maximally truth seeking. What was maybe a point in the past for you as a founder where you experienced that you weren't fully honest

Ilya Kravtsov: with yourself as a founder or as a CEO or, you know, you always, I mean, very often you're in a sales mode, right? In a sales mode. And you are also in an environment where everybody's in sales mode, right? So I think what is very important is to a certain degree not to oversell, right? And what you realize is that actually like overselling is always backfiring eventually. Right. Uh, and actually investors appreciate if you don't do that and if you even admit your problems is better. Right? So at the beginning when you start, you always want to not talk about your problems, right? Because you're kind of embarrassed and you are not confident. So I think the confidence also grows with time. Uh, when you start a first company, you're not confident. It's just like, oh my God, if I tell I have this problem, people will think that I'm a failure and all that, right? And I think it's very important that I would encourage all the founders to embrace the problems and be open about it, right? Because if investors see that, uh, you have problems means that you're first of all very honest, right? And that means that you also understand that you have a problem, right? And understanding you have a problem is very important, right? So I think uh, you know, that is something that also evolves. So I would always encourage any founders not to go and like, oversell, but, but admit you have problems, talk about it, ask for advice. Right. And then eventually solve them and then move on. Right. But it's very normal that any startup has a lot of issues, uh, as a young company. But that's totally fine.

Felix: Definitely. As an investor, if you get an email, an outreach, you know, this is our problems, this is how you could help. That's much more appreciated than receiving an email shortly before the funding runs out and you want to do a rescue round from your investment, invested founders. So definitely resonates with investors as well. I think earlier you mentioned the macro environment. Right. So I think Southeast Asia, um, when we just look at fintech funding since 2022, there's more than a 75 drop in volume and deal value. Are investors overreacting? Are there a lot of hidden gems now in the ecosystem that are not being seen? Or is this a fair correction?

Ilya Kravtsov: Mixed feelings, to be honest. I, uh, think, um, there's several aspects to consider. Right. So let's look at what happened between say, 2022 and now. Right. So one aspect is, uh, of course we went through a period of high interest rates that kind of like, um, was an issue for a lot of companies who wanted to go public. Right. And we haven't seen as many IPOs and exit in the region. Right. So I think that's definitely one driver for the slowdown in funding. I would say the second, uh, big drivers, uh, is that now we are in an AI era, which is completely different. Right. So, um, again, uh, um, you know, they say SaaS is dead. I don't think SaaS is dead, but if you talk to large enterprises and they need to spend their money, IT budgets, they will spend on the newest thing, which is AI. If you have SaaS in the back, it's very good because then your AI is just more sticky. Right. But like, everything evolves into the newest trends. Right. So I think, um, uh, so AI is out there. Uh, and then of course, selectively you can talk about certain countries in Southeast Asia with their own. Each of them has their own issues. Right. But, um, I think fundamentally what I believe is that we are entering more of a global era. A global era where you cannot just build for one single country or region. You need to build for the world. Ah. And that also, from an investor perspective, makes the investors, uh, shift their focus. Right. So maybe before you would raise your funds specifically for just deploying Indonesia or Deploy only in Southeast Asia. I think, uh, now this is evolving and investors will think that, you know, the newer funds and new LPs that would rise is more global. Right. So I think that is a big trend and that contributes, that the money flow goes away from specific countries to more of a global play. Right. Ah. And I think this is very fair. Right. And I think that's how a lot of founders also need to think. Right. So I always, when I mentor some younger founders, I also encourage them to think, oh, no, not which market you're going to launch first, but what is your top 100 clients around the world. Right. Think about your top 100 clients and doesn't matter where you are. Right. Because that's how the world is evolving. And I think naturally, uh, there will be great founders from Southeast Asia that will build from the world for the world. Right. We've seen some cases, of course, from Chinese founders are great at it. Right. Uh, and you know, they're doing so many, so much innovation. There will be more innovation from Singapore, etc. Etc. Right. Uh, but I think the decline in funding is not necessarily purely attributed to, uh, you know, stagnation or anything like that. It's just the world is evolving. You will see more global in vcs. Right. And even measuring it pure by country or by region, I think makes less sense.

Felix: And initially many funds looked at the Southeast Asian ecosystem. I don't think we need to talk about sort of broken expectations for some of the markets. I think when we looked at it, it was clear that Indonesia had the highest chance of, you know, delivering unicorns in certain spaces, at least out of Southeast Asia. Most other markets standalone didn't really have that opportunity. Singapore was always a stage for global founders that were strong in fintech, uh, partially in AI as well. We're talking as well about Chinese founders coming to Singapore, building great Web3 companies, or building great AI companies actually with amazing talent. Where does Indonesia now stand in terms of the opportunities? Right. Is there anything that, that you can compete, uh, and really build a right to win on a global level? Because what we see is, you know, many of the very large funds that raised lots, uh, of capital, they're either going to Australia, they diversify into Japan, ah, going to India. Right. If you don't have a person on the ground already, it's a very difficult endeav, or you go into more the PE place, which is basically funding, you know, mines or, um, food and beverage companies, which have very different economics. And that's also fine. Right. Maybe your LPs have signed up for something else. Um, but looking at Indonesia now, where, where are the plays that, that make a globally competitive business there?

Ilya Kravtsov: Honestly, I think Indonesia is a good um, you know, training ground. Right. And the reason why it's a good training ground is because um, it's a large country, you can get easily to high number of transactions. People are relatively open to work with tech companies. So even the more large enterprises, et cetera, are open to explore working with tech companies. Um, which is different than go to Europe, try to sign up banks to work with startups is very tough. Um, so I think uh, being able to like in the first phase of growth, like you know, testing and training and understanding and building your product in Indonesia is actually good. Right? Because you have a very cheap, you know, cost base which it's very important at the beginning. You can get to very large operational numbers. Right. Uh, which is also very important to perfect and iterate on your product. You can even get to you know, good brands in terms of like large enterprises that are present there that will work with you. Right. Where I would say it's more challenging is that the, the purchasing power is too low. Right. And that's why it's inevitable. Like if you are, you know, start the business there, you, you gain success, you establish your leadership there. It's very important, important to think on what's this, what's how this is going to evolve, right. And how do you leverage what you've built at the relatively cheap price elsewhere where building the same would cost you so much more. Right. So I think the, and by, so I think one is a training ground, two is by now of course Indonesia had a lot of like unicorns and you know, there is amount of talent that you can use to, at a relatively cheaper price to build this next generation product. So I would say uh, think about indonesia as your 0 to 1, but always have the next step planned. Okay. Where are you going afterwards? Right. And why do you need Indonesia? Are you building a product that requires a lot of training? Right. Are you building next AI that you really need to train of millions of data sets? Yeah. That's a good place, right. If you're building something that doesn't require that much of input and iterations, then maybe that's not the place, uh, you go to a market where your purchasing power is higher.

Felix: Yeah. Doesn't that advantage of labor cost arbitrage really go away very quickly once lots of the programming workflows. We talked about developers, right. A lot of this can now be replaced and 10x or 100x by cloud code and the tools that are going to come up in the next one or two years. So maybe on the developer side that advantage will completely disappear. So what it really comes down to in the end is localized relationships to bank partners. Maybe it's probably proprietary data and it's probably localization to other dimensions. Right? Um, what do founders need to look for? Where's the value that's really defensible actually in the end?

Ilya Kravtsov: I mean, just to give you a concrete example, right. We're still a relatively early stage company. I mean we have 10,000 customers per day applying for products on our platform, right? Right. Uh, getting to 10,000 customers per day applying for like a mortgage, right. In another country would be very tough, right? Because mortgage is not something that you do on a daily basis. Right? But on the other hand, having 10,000 customers applying for a mortgage per day, if you're building a good AI product and you really want to understand, you know, where's your dropout ratio? Like how do you do your proper eval models on your AI agent? And you know, is it if the agent answers, uh, too briefly, would the customer drop out and study all this? Uh, you need large amount of data, right? So like for example, I think this is for in our specific case becomes very valuable, right? Because we already, you know, uh, trained the models on a lot of iterations with a lot of customer data being still a relatively early stage company for a product which is mortgage, right. Uh, I mean some countries don't even have that many mortgages, right. Per year. Right. So I think that is valuable for sure. Now of course you could argue, you know, do you want, would you start it from like Singapore eventually or you started more developed market? I think that one is arguable, right? It's arguable. But I think the data side, the getting access to good enterprises early on, getting feedback, I would say more on the product side rather than engineering itself because product, you always need to talk to your customers and that's most valuable thing when you're building at the beginning, right. You need that feedback loop and that feedback loop in Indonesia specifically. It's easy to get.

Felix: You spend a lot of time in Indonesia. What's some of the misconceptions that uh, foreign founders have about that market?

Ilya Kravtsov: Let's map the world, right? So, uh, you know, we have us, of course, which is I would say the best market for AI, et cetera, the market with the highest purchasing power, but also the most competitive one, right. That's one. Right. Uh, we have, you know, uh, we have Europe, which is still a bit difficult, I would say. Right. Um, we have China and India, which are doing well, but are very different. Like very, I would say, closed environments that is hard to penetrate from. From outside. Right. And very close itself. And then you have, um, you know, Latin America, Middle east, uh, they're doing okay. Uh, Middle east specifically in the last two years, been growing very fast. Right. Um, but are, uh, not yet, of course, at the size of, you know, of China, India or and so on. Right. And then you have Southeast Asia. In Southeast Asia, Indonesia is by far still the largest economy. So you cannot ignore Indonesia if you are building out of Southeast Asia. Right. Um, so you cannot ignore it. Um, but the way you need to approach it is that, yeah, Indonesia, you will get a lot of operational tractions. You will not get to, uh, hundreds of millions of revenue that quick. Right. And that's why the mistakes on founders saying, oh, I have 100 million customers and this is amazing. Ah, but I'm not yet at, ah, uh, 10 billion in revenue. Right. So I think the misconception is that the market is. The purchasing power is low. Low. But getting to operational tractions is definitely possible because the people are really embracing tech. Right. Uh, so if you know that. Right. And then consciously you want to use, uh, this to your advantage, then go for it. If you think that that is not something that will give an advantage for you, you can focus elsewhere. Right. But that's what Indonesia has to offer at the moment.

Felix: Actually, you started your third company right after becoming your father. Father. Um, how did becoming a father change you as a person?

Ilya Kravtsov: And I know, yeah. So same for me. Yeah. I didn't know if I could disclose. Right. But so look, I. Yeah. So my daughter was born. Yeah. In the 19th of November, in 21. I started, incorporated the company the 7th of January, 2022. Right. Just basically one month and a half after, I think. Think a, uh, few thoughts there. Right. So, um, one thought is that, of course, family, especially kids, is a very good distraction. Right. So sometimes as a founder, you're like totally absorbed by your business and, you know, you cannot for, even on the weekends, you're always thinking about it. You're doing some Excels here and there. Uh, now Claude does it for you, but, you know, still.

Felix: Right, exactly.

Ilya Kravtsov: But with the kids, you cannot. Right. Because they're there. They require 100% of your attention. Right, Right. So first of all, I think that's a great thing. Right. Because it, it does distract Your mind from some of the things that you need, actually. Right. Second thing that you also start realizing is that, you know, uh, when you have a family, you have more responsibility, you have more financial burden as well. Right. So you also start pushing more timelines on yourself, which I think it's a good thing. Right. To get to a certain outcome faster. Right. Because you know that, that, uh, kids need to go to school. This. International schools are expensive. Right. So before, when you're a young founder, you feel that you have all the time of the world, right. Now you feel you are on a timeline and you need to make out of it the most you can, which is actually, you know, aligned with investors and their fund horizons. Right. So you want to really get to a point of, you know, am I growing 10x a year? Yes or no. Right. If I'm not, I need to do something very quick. Right. To get there. And I don't just sit and wait. Right. So I think that's the second. Second realization that I think gives things, you know, puts things into a different perspective. Right. For sure.

Felix: Yeah. And a different way. You're also compressing timelines on a, on a daily basis. Right. Um, similar to some of the accelerators compressing, you know, a few months of progress into a few weeks. Um, but. But it's kind of the same on a daily level for fatherhood, for sure. Maybe. To close the conversation, what's one piece of advice that you would give to

Ilya Kravtsov: new founders, especially since we're in Southeast Asia? I think, um, many founders come to me sometimes for like, oh, I'm doing a seed round, you know, can you look at my deck? What would you recommend?

Felix: Right.

Ilya Kravtsov: So I see few, few mistakes. Right. Uh, well, you can argue whether it's mistake, but few, like, comments that I have. Right. Usually one comment is, many founders are still stuck in the reality. Or I'll build a business first in Indonesia and then I will launch another market and launch market. Right. So they kind of like constrain themselves from the beginning.

Felix: Yeah.

Ilya Kravtsov: Which I think is wrong. Right. You should not think, oh, I'm going to build a business for, you know, Singapore first and then maybe from. For Indonesia, Philippines and then.

Felix: Yeah.

Ilya Kravtsov: Like, you should not be constrained by market. Right. You should think global day one. Right. I think that's, that's definitely first advice that I would give because the world has completely changed. Yeah. Second one is, um, you know, everybody puts AI everywhere. Right. And as we said before, like, when we went through that exercise of AI, we understood that AI can do something that like for example, selling advanced financial products that you know, the traditional SaaS could not do. Right. So it was a very clear reason why. Right now I see many founders putting AI for like OCR type of stuff that was available long ago. Right. So have a very clear idea of why you need AI. Right. Because if the mode is not very clear why you need it and you're putting it, it also dilutes the story a lot. So I think that's number two. And then um, number three is that if you are in a global space, you need to understand that you are in a hyper competitive space where the expectations on your growth are insane. Right. Projecting a hundred million in five years business plan like everybody used to before is not good enough.

Felix: Yeah.

Ilya Kravtsov: Right. So you really need to understand that if you are in global AI space, people go to 0 to 100 million IRR in like 10, 15 months. Right. Or even less so. So the level of growth and aggressiveness is very different. Right. So I think those are the three things that I see. People are still stuck a little bit in the past. Right. They start with smaller realities. They just put AI because it's sexy and then they like project, project, something like that is I would say conservative based on the assumptions. Right. But they don't think that big and don't think that in the eyes. At the end of the day if you don't grow at crazy rates, it's just going to be out.

Felix: Yeah. Harry stabbings at 20 VC said in the past it used to go 2 to 3x a year. Right. That's actually also the SaaS funding napkin, some of the common knowledge there. Um, now you need to be 10x or more. Does that mean 2 to 3x in Southeast Asia? You no longer, long longer of a chance to get funding?

Ilya Kravtsov: I think it depends. I think in Southeast Asia we see some VCs are um, you know, bridging the PE VC world. Right. So I would say many are going from like being a VC to more so a PE approach nowadays. Right. Uh, and if you're a PE, 2 or 3x is definitely okay as long as you, you know, you have a clear, you know, profit, you're profitable or close to profitable or have a clear fast path to profitability. Yeah. Right. So I think it really depends who you're talking to right now. Of course if you are growing 100x, you don't even think about margins, not even profitability. Right. So I think uh, it's really, you need to understand what type of business you're going to build and what investors you're going to appeal to and what is your personal ambition. Right. How big of an impact you want to create, how big you want this to be. And you know, the place where you started, it's also constrained by the investors there. Are there. If you start a company in Southeast Asia. Yeah. Maybe you need to think more about the bottom line. Right. If you start a company from Singapore for the world AI and you go to yc, et cetera. And this is different.

Felix: And as a consequence of that, if you're building a peace style business, you need to be comfortable being valued on the EBITDA level as opposed to.

Ilya Kravtsov: Of course. Of course.

Felix: Fantastic. I think that's all the time we have for now. Thanks so much for joining and see you guys next time.

Ilya Kravtsov: Thanks, Felix. Thank you.