He ran Prime fulfillment for Amazon—then raised $50M to replace e-commerce with AI. | Maju Kuruvilla, Founder of Spangle

Maju Kuruvilla (00:00:00) :
You start something small, and then you get distracted, and then your customers will start pulling you into directions where it is not aligned with where you want to go long term. And that's where you need to have high conviction of your vision. And we, every day, struggle with those kind of decisions as well. You're selling even before you're building, and you're selling even while you're building, and then you're selling even after you're building. So you're constantly, constantly doing it. I don't know whether the term product market fit is fully appropriate now. It made perfect sense in the SaaS world, where you build something, and now you can sell repeatedly for the next five years, ten years. I think that time or that phase of products probably is gone right now. Because now you have to find your relevancy every day, you know, you're one deployment of OpenAI or one Anthropic release away from dying.

Previous Guests (00:00:55) :
That's product market fit. Product market fit. Product market fit. I called it the product market fit question. Product market fit. Product market fit. Product market fit. Product market fit. I mean, the name of the show is Product Market Fit.

Pablo Srugo (00:01:07) :
Do you think the product market fit show has product market fit? Because if you do, then there's something you just have to do. You have to take out your phone. You have to leave the show five stars. It lets us reach more founders, and it lets us get better guests, thank you. Maju, welcome to the show, man.

Maju Kuruvilla (00:01:23) :
Thank you so much, great to be here.

Pablo Srugo (00:01:25) :
Excited to have you here, man. You've had quite a journey before starting Spangle. Tell me a bit about kind of, you know, you spent some time at Amazon, at Bolt, and then you started Spangle less than two years ago. So curious to kind of hear, your journey from your perspective and then we can kind of go from there.

Maju Kuruvilla (00:01:40) :
Yeah, absolutely. So I grew up in India, grew up in a small business family, loved computers and next thing you know, I was working for Microsoft here in the US. Spent a few years and then essentially kind of went into the startup journey at Amazon for a long time, then right before this at Bolt and started Spangle. And if you look at the journey overall, I spent a lot of time in commerce. Which is something I'm very, very excited about and want to continue to innovate around. But you will also find that I was always at the intersection of two things. Like at Amazon, it was e-commerce and the physical space. Before that, health care and payments. So I kind of always throw myself at the intersection of two functions where some innovation is ready to happen and excited to continue to innovate. And at Spangle, I kind of see the same thing.

Pablo Srugo (00:02:31) :
At Amazon you were a VP of technology but what exactly did you do? What did you work on?

Maju Kuruvilla (00:02:35) :
Yeah, so I had like two roles predominantly when I was at Amazon. One, I managed all the technology that used Amazon worldwide for powering Prime fulfillment. That means that anything you buy on Amazon, whether you get it same day or two days or even one hour. I was part of one of the teams that launched the one hour fulfillment back in the day, which was a crazy experience. But pretty much all the technology that we built to power fulfillment worldwide was something I managed. Just a crazy experience, fun. The opportunity to work with amazingly smart people, but also solve really, really complex problems at scale. But the second part of my Amazon journey was I became more of a business leader, like a general manager. I ran Amazon Global Logistics business units worldwide, which was interesting. When I was running Amazon Global Logistics, that's when the pandemic happened. So you had kind of a first row seat of watching how the world unfolded in reaction to the pandemic and then how to kind of manage through all of that. So had some great experience and learning both from an innovation side, scale, technology, as well as just running a very complex global business.

Pablo Srugo (00:03:52) :
And then, without going into the details too much but maybe just at a high level. What makes you decide to leave there and join Bolt? And what was kind of the, what was the idea at Bolt? What was the vision?

Maju Kuruvilla (00:04:01) :
I mean, at Amazon, I kind of saw the innovation and how it can be helpful for consumers as well as for a lot of sellers in the world. And the idea was that, OK, Amazon is fantastic. They are continuing to innovate and doing great things but it's all centralized in one consumer platform. So the idea of decentralizing and bringing that power to a lot more retailers made a lot of sense. And Bolt was doing that in a form of helping with checkout. So that was very appealing to me to kind of be part of that, providing that one click checkout worldwide.

Pablo Srugo (00:04:39) :
What were your roles at Bolt?

Maju Kuruvilla (00:04:41) :
I started as the CTO there and then became a COO and then eventually CEO, when I was there.

Pablo Srugo (00:04:46) :
Tell me a bit about what Spangle is and just the origin story of it.

Maju Kuruvilla (00:04:51) :
Yeah, so Spangle originated. Some of it is because I've been thinking about commerce for all the time, but a lot of it started after coming out of Bolt. I was speaking to a lot of people, one particularly who became my co-founder and CTO, Fei. Fei was the CTO at Saks OFF 5TH and he had the opportunity to run this large commerce platform by integrating a lot of vendors and a lot of pieces together to provide a great experience for the consumer. And so he was kind of thinking about how hard it is for all these brands to do this at scale, and how many different vendors you need to bring it together. And none of them really have a unified kind of a brain that's thinking what's good for everyone. So that was kind of the frustration he was kind of thinking about and he was a principal engineer at Amazon for twelve years. Before that, he built the Amazon consumer chatbot that a lot of people are using for customer support worldwide and also he was part of the Alexa founding team. So I knew him from Amazon. So we kind of started brainstorming some of his ideas, actually right here in this basement. We were thinking through that and the idea was that, OK, commerce might be a time where it might go through an evolution again, like e-commerce was a big deal. But maybe with AI coming in, this might be a very, very interesting time to go through a big transformation.

Pablo Srugo (00:06:23) :
Because this is when? This is early 2024 that you're having these conversations?

Maju Kuruvilla (00:06:26) :
Early 2024, where these things were just coming out and people, consumers, ChatGPT was kind of making a big difference. People were starting to get a feel for what AI could do, what it means, and consumer expectations are slowly starting to kind of evolve into. I mean, you know, every time the consumer directly can have a better experience than what enterprises and brands are providing, to me, that's an amazing time for innovating, right? Because that is a big delta. Because as a consumer, you can just ask ChatGPT and do a bunch of things. And when you go to an e-commerce site, it's kind of still a Frankenstein infrastructure, with a search put by some vendor and a recommendation by another vendor. And then all of these things cobbled together with static pages, and taxonomy structure that are traditional. Someone is pre-curating and so, when the AI and ChatGPT is showing that everything can be fully dynamic and everything can be dynamically constructed real time for you. Why the e-commerce infrastructure is not doing that and you can also think, that's the difference between a physical store and e-commerce, right? In a physical store, you can only structure it one way for everyone and everyone better walk through that. But e-commerce, you have an opportunity to literally create it for every single person exactly how they want it to look like. But we are not quite doing that. E-commerce is literally a copy of physical commerce with some variations. But again, it's all like a Frankenstein. So AI is kind of giving the promise that maybe this is the time to finally do what commerce always promised.

Pablo Srugo (00:08:07) :
Like a fully personalized commerce.

Maju Kuruvilla (00:08:10) :
Pretty much in real time and dynamic. I understand it's everything and if you and I can see, if I walk into a store thinking of buying a ski boot. Then it looks like a ski store and if I'm trying to come in, and buy, let's say, a soccer boot. It looks like a soccer shop and what if we can do that at infinite scale? And what if there's a technology that is powering that? And that was kind of the genesis of the conversation. And that is the genesis of Spangle.

Pablo Srugo (00:08:39) :
You know, I have this with a lot of founders sometimes. Which is that sometimes we're talking about e-commerce transforming and what does agentic commerce, and AI-enabled e-commerce look like, and all these kind of big ideas. And I think most people agree that it's going to change. Probably nobody knows exactly how, but one of the things is when the ideas are so big. One of the places that's hard, and I'm curious about your thinking on this. Is you kind of think to yourself, well, whatever it is, either Amazon's going to do it or Shopify is going to do it. You know, like how do you? Or ChatGPT is going to do it, right? It's going to be a platform where all the consumers already are or all the businesses already are that are obviously thinking about how AI is going to change everything, and they're going to just eat this up, right? When you're having these early conversations and you're very high level in terms of where you're starting. How do you break that down into something that, OK, this piece of the puzzle. This is something that we as a new startup could take on?

Maju Kuruvilla (00:09:34) :
I think that is one of the hardest things, rightly, you pointed out and we are seeing that a little bit, right? I mean, these models are coming out and basically killing each of the startups or ideas. Because they can do it faster and better, and they have more capital. But in some ways, that's not a new issue. We always had problems like that when, you know, Shopify arguably could build some of the features from their platform. Like their app ecosystem, and they are becoming popular, or Amazon can start selling some products that became very popular on their platform. So it always existed but now it's becoming harder and I think for us, what we did was I started calling a bunch of people who I know. For example, I spoke to Chris Rapp. She was running all the digital commerce for Victoria's Secret at that time. But she and I worked together at Amazon for a long time. She's a big innovator. So I called her and asked, hey, I'm starting to think about this idea of maybe AI should run commerce. And because consumers are probably humans today, maybe it's agents tomorrow or whatever. And also there is a million decision making happening. And I built, at Amazon, I built systems that can take enormous amounts of signals and make smart decisions at massive scale. And so, why can't we do this? Why can't we take all the signals in real time and then make great decisions, and provide something whether it's humans or agents in the future? And she was like, yeah, you need to really focus on one of the use cases to start with. And so the use case where we landed on was, where is that pain point that no one is solving for right now? And I found, through discussion with her and a few others. There is a gap today between how people are discovering products on marketing platforms and then how those people then get sent over to a digital site, and what is happening on that site. So for example, brands spend money to put an ad on Instagram. Because Instagram is so smart it can figure out who to show that ad to and let's say, Pablo, you see an ad for a soccer shoe, and you click on it, you are taken to a digital brand site that is selling that. The brand has no idea what inspired that consumer, maybe on Instagram. There is a story there, there was a reason, there is a context. There's something that inspired people and if you lose that context, you are restarting the experience again on a digital site. Whereas if you can bring that context through, you can make your site work as an extension of the ad that inspired you. So we started looking at that, and we found almost forty percent of the overall traffic coming to a commerce site today originates somewhere else. And usually it's some kind of a marketing, and that's evolving. I mean, it could be Instagram, it could be Google. Now it's ChatGPT, it could be agents, it could be browser agents, whatever. But today, as we speak, around forty percent of the overall traffic coming to a commerce site is originating somewhere else and there is a story that started somewhere else, which you are totally ignoring. So when almost half of your traffic starting somewhere else has a story, has a context, and if they are coming here. And if you are just completely ignoring all of that context, and trying to restart, guess what? It is going to impact a consumer experience. It is going to impact the trust, it's going to impact the conversion. Because you miss the opportunity to literally create a storefront for them that caters towards the context that they start. So we started looking in that specific gap to say, what if we can build our solution very tailored for this? And that's what we did. Our V1, our wedge product. We tell brands, hey, a lot of your traffic is originating somewhere else. You're spending a lot of marketing dollars to bring them to your site. Your conversion on those are pretty low. What if we can help you get more money from the existing money you're spending for all of these things without changing anything? Our AI can do that magic, and it can pretty much make you money by doing that. Would you like to start? And that's how we start. That is our V1, and it turns out it works really well. We have several enterprise brands now. All of them are seeing substantial results. We have case studies that demonstrate that we can bring up to almost fifty percent increase in conversion, thirty percent increase in revenue per visit. I mean, it's just practically free money that you can just take on without changing any of your existing infrastructure. That's where we started, and the beauty of that is it gives a taste of what is possible to the brands. Because now that they see that you could dynamically create experiences for people based on context, maybe we should do more of that. Maybe we can do more. So to answer your question, we kind of found that very, very specific problem and then really zeroed in on that. And made sure that we can deliver on it, and the results can be published as a case study so there is enough data to pay off.

Pablo Srugo (00:14:42) :
You know, I think it's. I have a bunch of very specific questions, but maybe at a high level. There's a lesson here in the sense that it's easy. Especially in these crazy times of change, like AI is really, you know. It is changing so many different things and some are clear the way they're going to go, and some are further out. But nobody's disputing that things are changing and it's easy as a founder who has many different ideas to get caught up in the end state, right? And kind of think to yourself, well, this is the way, things will be or should be. And then kind of getting stuck there. I mean, you have to somehow have maybe some opinion on that, but you also have to ground yourself in what is the current state of affairs and how can I make a big difference today? Not, in three or five years, assuming things go that way and I think that link is not easy. It's not easy to do, and I think sometimes you get lost. And sometimes you do find founders that are kind of building this future that maybe never happens or that's the big future that some big platform builds anyways, and so you never kind of get off the ground, so to speak.

Maju Kuruvilla (00:15:39) :
I think you're right and we ran into that at the very beginning as well, right? Because we wanted to build this AI for that run commerce and that's the holy grail here. But to build that to a point, that's a lot of work and you have to solve so many use cases. But the natural question would be, oh, why can't OpenAI do it or someone else do it? And they have more opportunity to do that. And this is kind of where a situation, where maybe you can build all of that. Or we said, let's keep that vision but break it down and let's start. What is the most useful thing you could build as the first step that is taking you towards that future state? And that's kind of how we evolved into what we look for. But there is also a danger in that, sometimes you start something small and then you get distracted. And then your customers will start pulling you into directions where it is not aligned with where you want to go long term. And that's something attention that you got to match. And that's where you need to have high conviction of your vision. And we, every day, struggle with those kind of decisions as well. So you're getting on a great point there, Pablo.

Pablo Srugo (00:16:52) :
And then just to get more clarity on kind of where the product sits. I mean, the first question I have is a simple one, which is obviously, the last five, ten years, we're going to call it landing pages and very specific landing pages that convert on specific marketing sets, ad sets. You know, have been the thing that works, right? Is this kind of the AI landing page, the next version of that? Because obviously, if you're running an ad set for soccer shoes, whoever clicks on that link is going to see the same page. Now, at least it's not like it used to be. Maybe where you just get directed to the homepage of your website or maybe just one section in your website. You're getting directed to a landing page that is very tied to that ad set. But like you said, I'm getting it, you're getting it, everybody that clicks that link is getting it and now what you're doing is you're kind of going that extra mile of specialization where we all might see a different landing page. Is that an accurate way of thinking about it?

Maju Kuruvilla (00:17:43) :
Yes, and one thing to add is traditionally, when you think of landing pages. They are manually created and also they are more visually oriented, like the text or where you press the button or the color of things. Which are, you know, interesting and important. But I am not a big fan of a lot of things because I view that as an agentic world. Agents don't care about the color of the button or the size or the creatives and things like that. What truly matters is merchandising, right? Finding the right product, the right categories, the most relevant things. How do you curate things together? How do you make sure that, using all the context of what I know about what you're looking for. Ideally, I want to show you the perfect products that I think you would buy without having to get lost in the transition. So it's less about the visual elements, it's more about merchandising and think of, you're an extremely smart merchandiser, like you walk into. Let's say, Nordstrom, and you're looking for a shoe for a wedding or whatever. And the idea is that Nordstrom can immediately present you with five shoes that you think perfectly work for you, or even show you ways to discover more things that are highly relevant to what you're looking for, or even asking you for more context so that it can further curate things. So yes, landing pages, but more focused on curation, merchandising, and also all the different elements that are truly functional. That are helpful for both the consumer and also the brand. Because you can imagine, behind the scenes instead of a pre-created template. An AI is delivering and rendering this experience. So the AI can have a dual mandate, where on one side it's delivering the great things for the consumer, and on the other side it's honoring what the brand or the retailer is trying to achieve. Which most of the time could be a conversion, but sometimes it could be acquisition, it could be, you know, pushing your brand or certain products, or giving more kind of guidelines around that. So take a step from the traditional kind of landing page, this is the AI running the landing pages. But think of it more like an entire store experience, not just a landing page.

Pablo Srugo (00:20:06) :
And then, another kind of product type question is if you're on mobile. Especially how much data can you really get from me? How much context can you get, in almost just how to get it? Because these are kind of think about an Instagram like it's pretty walled garden type situation.

Maju Kuruvilla (00:20:21) :
I think this is where I feel the consumers are always sharing their intent, with every click, every swipe, every activity and the question is, can you really take it all on and really make real-time decisions based on all of that? And let's just start with Instagram. I mean, you know what was the ad. Let's say the campaign and you know whether it's retargeting or prospecting. You know what was the creative, you know the copy, there is so much intelligence in there and if someone is coming through, you know it's a full-price ad. You know you are bundling some products together. You are curating for an event or a location or whatever it is. You have so much intent from right there. But on the other side, after they come in, there is a lot more intent you are coming into. You can see, after they come in, did they search for something? They clicked on something? Did they look for something? And you can real-time learn, adapt, and evolve into that. But if you are running it, let's say for a longer time. You kind of watch patterns of what is happening. People coming from certain ads, are they doing certain things? And this campaign, people are doing different things versus something else. And the model is always learning, it's adapting, it's evolving, it's optimizing. And that is one side of the world. Now, on the other side, we are also elevating the product intelligence as well. For example, you could take a soccer shoe and you can enrich that soccer shoe with all the different ways people are talking about it. All the different occasions those probably are happening, you know, whether it's a World Cup coming up or whether it's a particular player using those brands. There are so many different ways products are described, they are used, they are touched, and all the different ways. So you can also enrich your products in ways that the context can be built out from the product as well and then you connect this to both, real-time data and synthetic data. Now you have a really rich set of information and now you have an extremely powerful AI model that can decide what are the right things to show in the right way. And now, also considering some of the brand's mandate, you can bring it all together in real time. So there is no lack of context. The question is, can we listen, learn, adapt at a grander scale? And that's the kind of systems I have built before when I was at Amazon at a global scale. But now you can do that with more AI, which is the exciting part. Because before, you had to do a lot more mathematical models and all the other things. And now, actually with AI, you can do a lot of those things much faster, cheaper, and even more effective than ever before.

Pablo Srugo (00:23:17) :
You love this show, you don't want to miss the next episode. Why would you? So hit that follow button. Trust me, it's in your own best interest. Now, let's talk a little bit about that kind of big vision. Obviously, you just raised the $50 million Series A earlier this year. So obviously, this is the wedge product, as you said and there's many different places this might go. How do you see, I don't know about the end state. But over the next few years, what do you think we're likely to see in e-commerce? I mean, you know, today, if you think about a website. Whether it's Amazon or a Shopify website, doesn't really matter. We all kind of see the same thing. At the most, the product recommendations might change in the way that, you know, if we open YouTube. You and I will see the exact same UI with different videos, or Netflix, same thing. We'll see the same UI with different videos. Is that the future of how these e-commerce websites look, or is that still too close to today, right? And there's kind of a much more different version that is likely to happen.

Maju Kuruvilla (00:24:02) :
Yeah, I think my view is that it is going to be much more different than what it is right now for a couple of reasons. One, today we built all these experiences for the masses, right? Like I said, a physical store, you cannot design this for everyone. You just design what works for everybody and that's how the entire research happens. That's how you test things. That's how you deploy things. That's why we are seeing a lot of similarity, because that works. I mean, why are we constrained with that? There is no reason to be constrained. We could literally show everyone something extremely unique that matches them and in fact, they're going to be a continuous experience from where they are coming from. So, for example, you're coming from Instagram. Maybe your storefront should look just like an Instagram, but it's a much more expanded version. You put five products on Instagram and here you have five hundred products but it's like an extension. Or you're coming from Google, seeing a Google result. You come here, it looks different or maybe you're an agent coming in, and it turns out agents don't care about all the UX. They don't even want to see HTML. Maybe they want to see a schema, and maybe there are no pictures in that, right? And maybe there are no colors in that. And so now that could be very different, or maybe more content-oriented stuff. Maybe we can bring things together. So there is, I think, we have to unlearn or unthink about how the traditional websites are created. Which are template-based architecture with few optimizations here and there. I think that's probably either already dead or it's going to die and it's going to be a fully dynamic experience that is perfectly optimized for a future state with both humans and agents. Irrespective of whether people are coming directly there or indirectly through that and there might be a journey across different domains to kind of go through. And you need to win, even the marketing will change. Think about that, like today, how is the marketing? It's like, again, great images and the way you kind of discount, and adequate promotions are ways. Every site has a pop-up with some email. I mean, our agents incentivize the same way as humans because we built everything around human psychology. Agents may not have the same psychology, but they are not restricted by time, space, bandwidth, they're not emotional buyers. So how do you transform? So my view is that everything will transform in some form or fashion and it's going to be a mix of meeting everyone's need. Whether it's a human, then you need to meet them emotionally and if it's an agent, you need to be smart about winning the deal with an agent. Or maybe it's something in between and to do all of that, you cannot do that manually. You cannot have a team sitting together doing A-B tests constantly to find something. You need a smarter model on your side. My larger view is e-commerce will operate most likely with a seller agent and a buyer agent. A buyer agent is, it could be, I would define, maybe Instagram is acting as a buyer agent, right? It could be your own personal agent tomorrow. It could be your browser agent, or it could be Amazon working as your buyer agent. But whatever, there is going to be an agent who is responsible for shopping around and buying. And then, on the other side, you need an equally smart seller agent. That is every brand and retailer. They're not letting other people scrape their website and do all these crazy things by them. Why don't we respond? And so we are building to model that seller agent, right? So that means that it can understand who is coming in, where they are coming from, what they are interested in, and then the people who come from these channels kind of looking for these things. What are they interested in and what is the brand trying to do? And now, can we use all of that and be smart, and participate in this economy in a very different way? And it doesn't have to be, and I'm not a believer in saying everything is going to be one way or the other. I like optionality. I always look for, let's assume everything is going to happen equally, and what are the common kind of frameworks that you need to make it happen. And at the end of the day, you need an AI behind your store that can think for itself. Spangle's vision is to be that for you.

Pablo Srugo (00:28:19) :
You know, that point you made on optionality. I think is very important these days because I find, especially with these cycles, you get these false dichotomies that play out. I've heard, you know, and obviously everyone's talking about this on every podcast. But one of the narratives is kind of effectively, brands are dead, the stores are dead because everything is just agentic commerce and you're going to have this agent and it's just going to be hyper utility, right? Just like, oh, you need soccer shoes? Okay, let me go scour the entire earth, and OK, for you, exactly this. This is the exact shoe that you need and you know, there's a part of me that just kind of gets reminded of, wasn't e-commerce supposed to kill retail stores, right? You got a lot of retailers that are doing exceptionally well, not just Walmart but there's many. Something much smaller, like Dick's Sporting Goods. I mean, it's done exceptionally well over the last ten years as just a regular retailer and so I think that's an important point. Which is the future is probably going to be you're still going to have your kind of Instagram-type ads and discovery. You're probably still going to have a lot of people actually browsing a website or a website-like thing, whether it's personalized or not, and then you're probably going to now have this new thing. Which is agentic, where maybe you will ask it, "Hey, find me a pair of shoes," and it will go and do it. And it already does kind of a version of that today. Or maybe it'll also know you so well and it will recommend in the same way that Instagram kind of recommends. It'll say, "Hey, soccer season is starting, and I know last time you bought cleats was a year and a half ago. By the way, I found these ones." So, but I don't know that it's this or that. It's probably a situation of this and that, and we're just adding another kind of layer, and increasing the pie. I mean, I don't know if you agree with that, but that's kind of where my head goes.

Maju Kuruvilla (00:29:54) :
I agree one hundred percent. Because to your point, even with all the advancements we have seen. E-commerce is only between fifteen and twenty percent of the overall commerce. So that took twenty years or twenty-five years to get to fifteen to twenty percent of the overall commerce and agent e-commerce or whatever other form is going to take a percentage of that. But it's not going to be hundred percent of commerce and I know the brands, and retailers need to use all the different channels very effectively. And consumers will be using multiple modalities for whatever they are looking for. And there might be different categories of products, right? I kind of think everything I have on subscribe and save, maybe that's a good one for agenting. Today I have it subscribe and save on Amazon. Why am I limited to Amazon? How do I know? Ideally, I want my own personal agent. Which I think might happen in the future. I don't need to use someone else's agent, my own agent and I could say, hey, I want my shaving cream. Just always find the cheapest one for me. Go look everywhere. Go to Walmart, Amazon, and go find me. I don't care, I want this brand, I want the cheapest price by this day. But then there are going to be things like my shoe. I may be very particular. I'm not going to tell an agent to go buy me any shoe and then there might be even more special high-consideration purchases, and emotional purchases, and all of the above. So, even for people it may not be all the same and not everyone is going to be the same way. And the reality is the brands, I believe. To your point also, the ones successful are going to know how to approach in all these different modalities and channels. But again, it would be easier if you have an AI model that can always represent you everywhere and can always be ready to respond to a buyer agent, a human. Whether that's in a physical store, e-commerce, they're coming from Instagram, they're coming from Google search, they're coming directly. It doesn't matter if you can have a unified way of representing and then, also represent to win and do that dynamically. That kind of makes sense, which is why I think it is time for a new commerce, intelligent commerce infrastructure and that's why we were like, OK, it is absolutely time. Whether it's us or someone else is going to do it and it's not going to be an incremental improvement of Shopify's template or e-commerce architecture or Salesforce e-commerce platform. It is something you have to go back and build from the ground up. Where you have a brain, it understands everything, it takes all the signals, it can always dynamically provide something and that simplifies. If you can do that, that simplifies brands' operational load. I mean, you have several people trying to buy all the different pieces, trying to test them, cobble them together, trying to make it all, and again, it only works on your website. It doesn't work for this and that. If you can truly have a model that does all of that for you, and if that's smarter than everything else. That's perfect and I think the future is going to be great for brands. I think the future is going to be great for people with physical stores. You just need to participate.

Pablo Srugo (00:33:12) :
Tell me a bit then, just going back to the storyline. I'm curious who are the first few customers, maybe the first customer that you sign and how did you get them kind of over the line?

Maju Kuruvilla (00:33:20) :
Yeah, I mean, along these lines. I started calling people I know, I called Sam Biswas. He worked at 421, and I called Fiona Tan from Wayfair. All these people I had been friends with, had conversations before I started talking to all this and started asking, is that a way? I mean, as a founder, that's what you do, right? You're selling even before you're building, and you're selling even while you're building, and then you're selling even after you're building. So you're constantly, constantly doing it. So we wanted to start with a customer that can be relatively small but innovative, but also someone who we can partner deeply well with and our first customer was Uncline, and they are part of the WHP Group. You know, we talked to the leaders there, Gary Haas and Greg Schwend. They're both some of the smartest people in commerce. They understand, they really understood the problem. They were like, yep, we got a lot of traffic coming from marketing, and our conversion is not great. If you guys can help us get more revenue from that, we're all game for that. So that was our first customer, and we worked together with them to learn and had almost every week kind of go through the progress and all the different pieces that we are working with the metrics. We want to agree on what metrics we will agree on to work with, and we were able to deploy very quickly for them and were able to demonstrate great results with them. And also got a case study with them together. And they have been fantastic partners for us. And that literally was our first customer.

Pablo Srugo (00:35:04) :
How did you set it up? Was there kind of a pilot phase? Did they pay for it? Did you kind of preset what the ACV would be like? How did you structure all that?

Maju Kuruvilla (00:35:11) :
We gave a free POC, being a great partner as an innovation partner with us. But right after, we made some of the metrics. Initially, it was a ref share model. We wanted to make sure that our pricing reflects a joint success, and we want to be incentivized to drive more business to a brand. And the brands should be happy that we are working because we only get paid if we are doing good things.

Pablo Srugo (00:35:38) :
Is that still the pricing model for ref share?

Maju Kuruvilla (00:35:40) :
Yeah, largely that is the pricing model. That largely works. I mean, enterprise deals would kind of go here and there, but largely, it's a success-based pricing with ref share.

Pablo Srugo (00:35:49) :
And revenue-wise? You guys are what? Kind of in the seven-figure ARR sort of range?

Maju Kuruvilla (00:35:52) :
Yeah, we're not sharing some of the numbers. Unfortunately but yes, we came out of stealth, I would say, less than a year ago. So we initially started, we spent a few months building a model and training, and making sure that all the pieces are ready. So we only came out of stealth around ten months ago and since then. We target mid- to large enterprises, and we started with apparel brands initially. So we have eleven enterprise brands so far, and the traction has been good. We have been limiting also, we want to make sure that we deliver really well for every single customer of ours. We're not trying to get a massive number of customers. Our approach is more of a partnership approach, where we get mid-level enterprises, and you work closely with them. We can demonstrate one use case, and then usually our customers really like what they see, and then, they expand with us for more and more use cases. As I mentioned, our secret sauce is our AI model. We call it Product GPT, which is the AI model. There are several use cases that can be built, but we have a wedge initial use case that's very simple. That can quickly demonstrate results without a lot of effort from anyone and so we kind of start from there, and then people want to use our AI for more, and more of their functions. With the view that AI can practically run the store long term.

Pablo Srugo (00:37:17) :
Would you say that you found true product market fit?

Maju Kuruvilla (00:37:19) :
That's an interesting question. We can finally talk about it. I don't know whether the term product market fit is fully appropriate now. I'm just thinking out loud here because it made perfect sense in the SaaS world. Where you build something and now you can sell repeatedly for the next five years, ten years. I think that time or that phase of products probably is gone right now. Because now you have to find your relevancy every day. You know, you're one deployment of OpenAI or one Anthropic release away from dying. So you cannot sit back on the laurels of, hey I achieved product market that I'm going to scale. In a traditional term, I think we have, but I am a paranoid person. So, I'm thinking, OK, I think it's a sliding scale. Which is always a moving target and also, as I mentioned, our goal is much bigger. I would say we will achieve product market fit when we can run an entire store on our AI, but we're far from it. But if the question is, is our Verge product work out of the box and is it repeatable? The answer is absolutely yes. We have several case studies, great brands, great references, and we have more demand than we can handle, which is why we raised the Seed. So we can kind of move a little bit faster. But the whole question of product market fit is an interesting one to kind of think about. I'd love to hear your thoughts on that. You are obviously an expert in this and talk to a lot more people.

Pablo Srugo (00:38:52) :
I think it's a fair point. You know, I think the reality of product market fit is it's always been a spectrum, and there's always been almost a part of that spectrum where it kind of clicks on. And, you know, before that you don't have it, and at some point you do have it. But it doesn't mean you can't keep closing in on it, and it certainly doesn't mean you can't lose it. What has changed is because things are evolving so fast, because products are launching so quickly, and the capabilities underpinning all these products are changing literally on a weekly, monthly basis. That the amount of time that you can hold on, kind of to your point, to product market fit has shrunk. It is just a faster timescale, right? So it was before you could get to somewhere and say you have product market fit. Yes. The market will change, competition will change, things will change, so I need to keep adapting. The level of adaptation was not as high as it needs to be today. Because you might have a product that is on the frontier, insane capabilities, and in three months, those capabilities are like status quo, table stakes, and nobody cares anymore. And so you can't rest. I think the word of the day, or of the year, let's say, is speed. I think that's the key variable these days as a founder; you just have to be so fast, and on the one hand, it's exhausting, on the other hand, the opportunities are bigger than they've ever been. And then me, so as a last question, what would be your number one piece of advice for an early stage founder that's kind of in this early stage?

Maju Kuruvilla (00:40:10) :
I mean, one I think about always is the same one everyone has been talking about. Which is just get started. Because everything is changing. So if you want to really sit back and truly find what exactly you're building. That will change by the time you even start building, right? So if you have a conviction about some kind of a future state or a big problem that you feel you want to put your teeth into and really solve. So let's get started. Don't let anyone tell you that someone at OpenAI is going to do it, or someone large is going to do it, or some other competition is going to do it. I mean, everyone is going to do it and you just get started. You work closely with the customers. What I'm proud of, a few things, is we are the best team. So having great people around you is absolutely key. You don't need a big team, and probably you should not have a big team. You should have an extremely small lead main team, and then you should have great customers. You should handpick your customers. People say you should handpick your investors, which is also true. Which I thankfully have done, but you should also really be careful who are your customers and that is more on an enterprise play. Maybe if you're building a consumer product, that is not relevant advice, but you've got to be very thoughtful about who are your early customers. They need to be your partners, and, you know, just get started.

Pablo Srugo (00:41:32) :
Perfect. Well, Maju, thanks so much for jumping on the show, man. It's been great.

Maju Kuruvilla (00:41:35) :
Fantastic, appreciate you having me.

Pablo Srugo (00:41:37) :
Wow, what an episode. You're probably in awe. You're in absolute shock. You're like, that helped me so much. So guess what? Now it's your turn to help someone else. Share the episode in the WhatsApp group you have with founders. Share it on that Slack channel. Send it to your founder friends and help them out. Trust me, they will love you for it.