Simile raises $200M at $2B valuation to build foundational models of human behavior for Fortune 500 retailers and banks

Speaker 1: And our next guest is Joon Sung Park from Similee. He's back. Co founder and CEO with a huge fundraise.

Speaker 2: Joon, how are you doing? Welcome back to the show. Hi, everyone. Great to Great see you. It's been Dude, you've been cooking. You've been cooking. It must be

Speaker 1: must be about a year since you're I don't know. It's blown it's blown by, but congratulations. Why don't you give us the news first because it's the first Gong hit of the show. I gotta warm up the Gong, and then you gotta tell me what happened. How much did you raise?

Speaker 6: We raised $200,000,000 at 2,000,000 Thank

Speaker 3: you. So

Speaker 1: where is the company today? Take me through the footprint. How big is the company? What really unlocked this new round?

Speaker 6: So the last time I was at TPN, I believe it was about five months ago. Since then, the company, literally the team itself, has quadrupled. And one of the amazing things that we're seeing here is the pure market demand. The SIMILI is a company, it's an applied AI lab that is creating foundational model of human behavior, where we create models that can predict human behaviors across different market segments in the future and the multi agent simulations. And this particular area of study and model is seeing an extreme amount of demand across retail. So we've been working together very closely with CVS, but other Fortune 10, Fortune 50 retailers, but also in finance, in some of the largest financial banks that are out there today, and also in CPG companies and more. So the market demand has been incredible.

Speaker 2: Demand for AI products is insane everywhere, And, clearly no different here. How are you grading yourself? Because it sounds like this this feels like the product, the kind of product where like the value prop, if the product works is like almost too good to be true. Right? Try to understand people's future actions so that you can better serve them as customers and all that kind of thing. And I feel like humans want to believe that, you know, they have free will and like, you know, you couldn't possibly predict my next move, June. I'm too unpredictable. I'm sure you disagree. But like how good how good is the product today and how much better can it get? Like from a from an accuracy standpoint and and and on that, like what what gets a CVS or one of these big logos confident in your product where they're actually changing their roadmap based around your data?

Speaker 6: That is a great question. So there's a technical side and the market side. On the technical side, today we may have already seen many models that are trying to be an amazing reasoning model. So many of the labs are working on this. So they would go to more core skills of the world, get expert data in coding, natural sciences, mathematics. These are trying to solve the objective problems in our lives. Similarly, it doesn't actually care about any of that. The models that we create are models that are trying to be as human as possible to actually represent the human values, preferences, taste, all the subjective half of human brain. So what we do is we get into partnership with places like Gala of the world that are amazing at creating a representative sample of human population. And we try to understand what is the real behavior and distribution of our human population. So really on the technical side, the success here is can we predict that distribution? Can we predict people's behaviors? And can we represent them at the scale that we want? The vision here is to actually represent only a billion people along the way. And we're very much going aggressively towards that momentum. And right now, we are representing tens of millions of people as we speak. Yeah. So that's the technological side.

Speaker 2: How much yeah. Yeah. Continue. Continue. I have another question, but continue.

Speaker 6: And, of course, on the market side, this technology is now having real impact. So we've been in deployment. So we're obviously a fairly young company. However, the simulation as technology has been in deployments on the in some of the largest markets in the world. And companies are making real decisions. They are usually starting from actually replicating what they know to be ground truth. The studies they have run, the behaviors of their customer they are aware of. Now we run them in simulation. We see that they replicate with extreme accuracy. And once they replicate, we go on to test new markets to basically help them test pre deployment testing on new markets, new product, message testing, and actually show ROI on those predictions.

Speaker 2: Okay. Very, very cool tool. Sounds like it's working. This feels like the there's there's one scenario here. The the sort of sad scenario is, like, big companies get access to the crystal ball that allows them to just, like, compound their scale and their growth and their customer bases, and and they have access. You're you're building, like, predictive super intelligence or predictive intelligence. The biggest companies in the world have access to this. Meanwhile, the like long tail of, you know, Shopify brands and smaller companies are just stuck, you know, guessing. I'm super curious. Do you think there's a product to be built that, you know, the the long tail of companies can can leverage to get access to that same intelligence? Because I feel like a big company can afford to spend $20,000,000 launching a new product and, like, swinging and missing. So they have the advantage of scale and being able to take different shots on goal and they have this sort of like base of successful products. Whereas you have a young startup that's just trying to get off the ground and you have maybe one, two shots to make something, you know, thinking about consumer brands. If you're a consumer brand and your first or second product aren't a hit, like you're probably not around to launch the third. So I'm curious about how many people you can get this to.

Speaker 6: No. It's a great question. I would actually look at this as the way to democratize access to people. If you look at how Insights really operates today, if you're a large enough company, you have the budget, and you have access to your customers to go after those customers to better understand them, to actually talk to them, get their feedback. But especially if they're a younger company that is small, that doesn't have a big budget, that is very cost prohibitive in many ways. Simulation is the way to get access to people in a scalable way. So it is actually a way to democratize this function. You can also think about this from the people's perspective, people who are represented by simulations. There are so many decisions that are being made today where we would love to be able to listen to people and actually see and get their feedback. However, in practice, it's very difficult. You cannot actually go talk to millions of people every time you need to make a decision. But if you can create simulations of people, then that is representation at scale. So in the grandest sort of scheme of simile, what we are really trying to achieve is to bring human voices to the rooms where the most important decisions are being made for our society.

Speaker 2: And what does that actually look like in practice? Is that like people are developing a product, again, I'll use like a big CPG company. They're developing a product and they can actually like pitch, like effectively like like I'm wondering how much this is, like, text based right now versus actually conversational. Like, can you simulate talking to your customers? I'm sure the YC folks would laugh at this. Like, stop simulating talking to your customers and just talk to your actual customers. But it but it feels like very valuable. You know, there's there's every founder's had the experience of like having a big customer meeting and it either going great or not going great. And sometimes you wish you could have like just done a bunch of simulations of that conversation before you actually went into it. So what is this actually like, how are people interacting with the product now, and where will they and and what are the other kind of interaction formats of the future?

Speaker 6: Really, right. So our customers come in, and what they see is an interface that allows them to filter down to a population of their interest. So they can describe the population to be whatever they want it to be. Let's say, male living in California in their thirties. And then they can literally ask any questions. That can be behavioral

Speaker 2: environmental questions. What about male podcasters in their 30s in California?

Speaker 6: You know, you must have guessed. You know, I simulated both of you before coming on today. You did it last time. Just like last time. Like, we've had this conversation thousands

Speaker 2: of This

Speaker 6: is my thousand first time I'm having this conversation with you all. But our customers can literally filter down to any population of their interest and they can ask questions. It can be survey form. They can also send images, if it's image asset, videos, if it's a video like advertisement, it can even be a product demo, whether it's a Figma or real website. And our agents will actually traverse through those websites and Figma mocks and basically give feedback once they have seen it. Now that's how the product is being leveraged today. But one of the things that I also find to be quite exciting is as agent take development goes much more prevalent, the cost of production is going down every single day. Now, the real alpha in that case really is to understand what do people actually care about? We can generate 10,000 variations of this product. Which one actually matters to people? So in that way, SIMILI also has an opportunity now to actually inform not just human decision makers, but but also agentic decision makers who have delegated power from real humans. So this is done through MCP, API. These are the kind of products that are also getting built instantly.

Speaker 1: Mhmm. Yeah. Oh, okay. Very interesting. Thank you so much for coming on the show. Congratulations on the progress. Really, really awesome. Yeah. One of the most I love how sci fi the company is, but also grounded in in practical business Yeah. Objectives. That's a great way to put it. Yeah. It's like very, very sci fi, but like then it's like CVS is a big customer driving actual value, which is great. Next time you come on, I'm sure it'll be soon given given

Speaker 2: Trajectory. Given the trajectory. But I want you before the conversation to predict, you know, 10 questions that you think John would ask, 10 that I would ask, and then and then let's let's compare at the end. Nice job. I

Speaker 1: have a feeling you'll be able to, like, get Probably at least, like, 60% accuracy. Yeah. But this is great. Well, thank you so much for coming on the show. See you soon. Congratulations. Cheers. Talk to you soon. Let me tell you about Shopify. Shopify is the commerce platform to grocery business and like you sell in seconds online, in store, on mobile, on social, and marketplaces, and now with AI agents.