She raised $50M by breaking the VC Rulebook | Larissa Schneider, Unframe AI

Amit: Today we have Larissa, who is the founder and CEO of, you know, this very interesting company. I would like to call it the new age Palantir.

Larissa: Over the course of a year, roughly, we raised two rounds—seed and then Series A—raised $50 million from VCs like Bessemer and Craft. We realized that all these AI use cases that come from different industries, different departments, different teams within an org, the actual technical components that you need to solve them are actually pretty similar. Think about: we built a whole huge box of Lego bricks that now we can use, collect, adapt, modify to specific customer use cases that they bring to us.

Amit: Why should I work with somebody like you? What is the advantage that I can get?

Larissa: For many of the use cases, you can maybe find something off the shelf, but that means you are using a product that already exists for the masses. Like, think about it in this mindset. If you talk to a Wall Street bank right now, they tell me all the time that they have a backlog of thousands of use cases. They cannot evaluate individual point solutions for every single one of them. That is not a scalable approach. In many enterprises, people were just trying to use ChatGPT or Claude to do something which was very specific, and you needed something custom to be built for it, but they were just trying to use ChatGPT to do it. Your employees will do something with AI if you want it or not. And in the worst case, they will put your company data in their personal ChatGPT subscription.

Amit: Where are you seeing the most amount of adoption?

Larissa: Software development is definitely one of the big things anyway. So I would say that is definitely the number one. I'm happy for our top use cases maybe not being in like the super sexy, shiny areas. I love the use cases where enterprise leaders tell us like, "Wow, this is a problem and it has been for like 10 years and we've been trying all kinds of different solutions and we haven't been able to. Wonder if you guys could do this."

Amit: For the longest time, VCs would believe that we'll invest in companies which are very focused on one thing. Basically, you find out a very narrow problem and try to solve it and stay in that lane for as long as you can till you do like 5 or $10 million and then you start expanding. But what you are saying is that you have completely changed that model and you have been successfully able to raise $50 million as well. How did you do that and what is the mantra behind it?

Larissa: Conversations with no second that your arms don't stop to think.

Trailer Ends

Full Interview

Amit: Hello everyone, to yet another episode of Build AI. Today we have Larissa, who is the founder and CEO of, you know, this very interesting company. I would like to call it the new age Palantir. Larissa, would it be a fair description?

Larissa: I mean, um, I'll take it, and I can walk you through a lot of the differences. There are a lot, but, uh, it's definitely an interesting description for sure.

Amit: Right. So tell us, like, why is it called Unframe AI and what is the origins? When did you start this company? How did you start it?

Larissa: Yeah, absolutely. So, um, I'm one of the co-founders and COO. Actually, all three of us were co-founders. We started the company at the beginning of 2024. We previously met in the cyber security space where we're all pretty active. Shai, who's our CEO at Unframe, he previously built a company called NoName Security, where he was the founder and CTO, a leader in the API security space. And while we were there, we realized, like, that LLM moment, the ChatGPT moment, happened and it was super, super exciting and really changing everything that we were doing in our personal lives. But the business side of things wasn't quite there yet. You were like still going to work and working with really old, legacy, clunky software every single day. And so we said, this is a fantastic moment to jump into AI and do something there.

So we, um, over the course of a year, roughly, we raised, uh, two rounds—Series, well, seed and then Series A—raised $50 million from VCs like Bessemer and Craft, and built a team of close to 100 people right now. Also, um, been, uh, interviewing a lot, been selling a lot, been raising a lot, but it's extremely exciting.

And you asked about the name, Unframe. Actually, you know, we started as a multi-product, multi-persona, multi-industry company from day one, which is not the normal thing to do in business. Like, often you start very narrow and then you, um, expand from there. We turned all of that on its head from day one. And, um, we're organizing our different solutions actually in what we called "frames." Like, okay, let's let's break the platform out of those individual frames and build like one cohesive story on top of it. So we were like unframing it, and, uh, that's how we started.

Amit: Fascinating. Now, as a as a builder ourselves in the AI space, I have a lot of questions based on what you just said, right? So let me start with the first one. I have invested in about 32 companies and partnered with a lot of VCs and investors while we are doing that round. For the longest time, uh, VCs would believe that, uh, we'll invest in companies which are very focused on one thing. And, you know, basically you find out a very narrow problem and try to solve it and stay in that lane for as long as you can till you do like 5 or $10 million and then you start expanding. But what you are saying is that you have completely changed that model and you have been successfully able to raise $50 million as well. How did you do that and what is the mantra behind it?

Larissa: I mean, I can also tell you, we got a lot of nos. Because a lot of VCs have their boxes in their investment memo that they have to tick, and we just didn't fit a box. So, that happened. It's not like it was always just a walk in the park and super easy to raise. Like, let's, let's set that aside. But we really, truly believe that our platform is so strong because it knew from the first moment of it being created, all kinds of different departments in an organization, industries, verticals, different workflows that are completely different across industries. And that really made it a lot stronger, and we would not have such a solid tech and go-to-market motion, even at this stage, if it wasn't for that decision very early on. Plus, this is the first time ever in history you would have been able to do this because AI has so many different opportunities that it opens. It really helped us, and in a way, we're super lucky that we started this company when the LLM was already out in the market, because you wouldn't have been able to do this before.

Amit: Right. So you are talking about early '24. So ChatGPT got launched in November '22. So basically there was like 14 months of LLM-based chatbots and things around but not really much. So like minus one day, they, minus one. Yeah. How did this idea come into mind that we'll build an AI-first company, and that too which is not like a single product? But I'm guessing, if you can also throw light on, is it a set of products? Is it a set of services? Is it somewhere in between?

Larissa: Somewhere in between, all of it. But think about it. We realized that all these AI use cases that come from different industries, different departments, different teams within an org, the actual technical components that you need to solve them are actually pretty similar. So what we did is we broke them down into individual building blocks. We call them almost like Lego bricks. Think about: we built a whole huge box of Lego bricks that now we can use, collect, adapt, modify to specific customer use cases that they bring to us. And then we tailor them and configure them towards a specific use case. Very, very custom to the customer. And we do that with something that we call a blueprint. The blueprint, you can think about, is like the Lego instructions booklet that you see in each of the boxes of Legos. And that's really what tailors the solution and makes it adaptable to the customer. And so, as we get to meet new customers and more projects that we work on, that's how our Lego box gets bigger and bigger. And that's why we're able to do so many different customer projects, right?

Amit: This is super interesting. Actually, in a way, we have a very similar thesis. So, we are a venture builder, and what we do is we help domain experts build their companies. So, we have also built a bunch of sort of, you know, Lego pieces, like our own agentic orchestration system, you know, our own document processing system, our own, you know, XYZ without going into details. And then it helps us every time that we are building a vertical AI company with the domain experts. So, I completely get it. In fact, I'm, I'm like, very happy to hear what you guys are doing, which is very, is like validation almost. It validates the market every time to hear someone who does something slightly similar, and, and yet, you know, just to sort of evaluate it from, for the benefit of the audience: that let's say if I'm an enterprise in financial services or commerce in the US, and I'm actually evaluating a product which is custom-built for, for me, like let's say I'm a lending firm and somebody has built a lending AI solution. Why should I work with somebody like you? What is the advantage that I can get?

Larissa: Yeah, I mean, look, for many of the use cases, you can maybe find something off the shelf, but that means you are using a product that already exists for the masses that was probably generated for the first two or three customers that that specific software company had. Now, this might not fit your processes, your workflows, your needs to a T. Like, you have to now turn your internal processes on their head to fit the software. And what we believe is, like, right now we have the power to give everyone what they're actually looking for, what they need. Not have the software determine how you run your business, but have the software fit how you do already run your business. That's super important here.

Also, a big advantage of working with, not a point solution—like, think about it in this mindset. If you talk to a Wall Street bank right now, they tell me all the time that they have a backlog of thousands, hundreds of use cases. Like many, many thousands in most cases. And they cannot evaluate individual point solutions for every single one of them. That is not a scalable approach to source, procure, govern, maintain all of those individual solutions. Impossible. So what you want to do is find a scalable approach. And that's why it's interesting to work with a platform like ours. This kind of goes back a little bit to, like, the new age Palantir model. We have a shared layer that all the solutions run on. It's called the knowledge fabric. And, uh, every time you work with us on one solution, and now you have another use case, another pain point within your company, we already know a lot about you. Like, our platform is already versed in your terminology and your data sets. And just that, those learnings, it creates this whole fabric, and it really helps you to onboard new solutions, new use cases onto our platform, and then being accurate, tailored, and effective from day one.

Amit: Right. And, and actually, I, I'm probably picking up the signal here, but, uh, for the benefit of the audience, Palantir has made this one term very, very popular, which is called FDEs, like forward-deployed engineers. And I think it's quite popular in the valley right now. But in your model, are there FDEs, or is it something very different?

Larissa: Absolutely not. And maybe this is not the most popular opinion, but honestly, I believe that you probably only need FDEs if you don't have the technical capabilities out of the box. Um, we build our tool, our platform, our set of solutions in a way that we don't need it, because it's really cumbersome, actually. Like, in most cases, an FDE is deployed for, like, 12 months, 18 months, something like that. It's a lot of onboarding, a lot of permissions within the organization, horrendous amount of cost. So, we purposefully did not want our customers having to do that, because you don't want to have, like, a very strategic decision of spending, like, tens of millions on an AI experiment. Like, that's in most cases not going to fly. You want to have, like, a whole transformation, but do it step by step, like small wins, quick wins. That's usually the best way of showing value.

So what we have is, we literally jump on a call with our customers for, like, 45 minutes, 60 minutes, um, and we really understand the first use case. Like, where is a good starting point? Where should we, um, tackle this partnership? And we bring in one of our AI product, um, leaders. They listen to it. They know how to ask really, really good questions. And usually we're able to get a great understanding of the use case. We go back for about five days, and then we meet the customer again, and we will show them how we translated this pain point, this use case, into a solution. It's ready for them to try. So it is really not just a mock or prototype. It's a functioning solution that they can play around with. They can either use dummy data, they can hook up to their real systems, and they can see if they get value. If they don't, no problem. If they do, in the next step, we'll move to, um, licensing, and then we charge per solution per year. Um, really, you know, how much you will be spending. There's no gotchas, no surprises after the fact, and absolutely no cost or commitment before you actually feel like you're really happy with the outcome.

Amit: So you don't charge anything up front, and you don't charge anything like a monthly retainer or anything like that. And then you just, once the solution is ready, the customer starts using it, they're happy, then they start paying for it. And, and what's the business model after that? How do you charge?

Larissa: It's, uh, it's normal pricing per solution per year. So t-shirt based: small, medium, large, extra large, kind of based on the complexity of that specific solution, and it's an annual, annual subscription. So if after a few years you feel like, okay, I'm not getting the value anymore, no harm done. You can stop the subscription, rip it out anytime, right?

Amit: And who owns the IP, in, in, in this case, uh, for each of your customers?

Larissa: Yeah. So the IP for the platform is ours, and it has to be ours, otherwise this business model wouldn't work, right? But every single blueprint that we work on with the customer—so think about it as like a compilation of, um, different building blocks that we, uh, tailor towards them—that will only ever be used once for that specific customer. So there won't be multiple companies running the exact same blueprint. That's, that's not the case.

Amit: Okay. While we have to own the IP, it's a, it's a single-use type of, uh, engagement, right?

Larissa: So it's fascinating. Like, before AI, this would have been very difficult to do. You're almost like a new Infosys or a new IBM, and all of this customization at mass scale was just not possible. But I can totally understand because we are doing something very similar, more for building these vertical AI products. How is it possible? So it's, it's a fascinating time.

Can you tell us a little bit also about, like, when you started, who were your first few customers, and today are these predominantly enterprises, and how has it changed over a period of time?

Larissa: Yeah, absolutely. It actually, uh, is predominantly enterprises, because that is honestly, that's our background. That's our network. All of us have always been in enterprise-selling companies. So that was what we know, and that's where we went after straight away. Very, very quickly, within the first few weeks of the company, um's existence, we went to our first conference as well. So I think we were week three, uh, when we went and exhibited at a, at a conference and started meeting people. And we found very quickly that financial services was a huge draw, because there's so much pressure, like from the board. It's the top one priority to compete against everyone else in the business. So that was an, an early win for us. Insurance as well, lots of unstructured data. That's how they came to us. Automated claims processing and so on. And real estate, that's also a big market. In the meanwhile, we work across many, many industries. Uh, but I would say the, the three—plus manufacturing, plus, uh, consumer goods—I would say are the, are the ones that, uh, we see coming back and back over and over again.

Amit: We couldn't agree more, because financial services is also one of the major—actually, you know, I built a fintech company before, and my CTO also built, was a CTO of a fintech company, and our third partner, who is based out of Bay Area, he actually was in the ventures team of BBVA, BBVA Ventures in San Francisco. So we have a lot of financial services experience, and it's very good to hear that that is turning out to be one of the biggest verticals as well from your experience as well.

Larissa: Absolutely. Absolutely.

Amit: I want to talk a little bit also about your own personal journey, um, you know, across companies. Like, you know, if you, if you can tell us like, u, you know, what are the companies you worked for, and then how did that shape up your experience and skill set to be able to do this today?

Larissa: Yeah, absolutely. So when I first set out, uh, like, I always knew I was going to be in the business world: marketing, uh, go-to-market, partnerships, uh, entrepreneurship. Like, those were the things that I studied in school. But I thought I would apply this to the biggest companies in the world. So I wanted to work for, like, the, the big brands. I, I went out to, uh, work for Airbus for a while, which I thought was absolutely fascinating, and I learned so much, and it was really cool to, like, see your product flying in the skies every day. But in go-to-market and marketing, I was like, you know, my next product launch would be in around 30 years from now. And this was like, kind of, I felt a bit suffocated by the, the speed in, in large enterprise, especially in something like, you know, air travel. Um, it's just not very dynamic in, in, like, the launching new product side of things.

So, I tried to do the exact opposite to that, and I, um, wanted to move to the Bay Area, because I'm originally from Europe, living abroad for, for quite a few years. But I was like, okay, what's the opposite of what I was doing? And that was moving to, to San Francisco and study startups. The best way of doing that was to go in there with a student visa. So, I took my masters in San Francisco and then really looked for really innovative new companies, uh, in the Bay Area. Found, um, a really cool company. I, I loved working there. Doesn't, isn't around anymore, called PernixData in, in South Bay. Started working there, and they were looking to open an office in either Berlin or Paris at that moment. And I was like, "Okay, let's, let's go. I'm originally German." They moved me over. We started the office here, uh, in Berlin, where 10-plus years later I'm still today. Oh, you know how it goes. Like, you somehow get stuck there. But it was, it was really cool. I, I loved that experience. And we're like 100-ish people. Within a year, we got acquired by Nutanix, also still San Jose-based, lots of presence, uh, around the world, big offices also in, in India. And, um, that we IPOed, which was another really, really cool experience. But I joined them when they were around a thousand people and left when it was, like, I think close to, like, 6,000, 7,000, so, uh, it was, uh, a completely different experience, growing up with that company, if you will. Um, did a lot of acquisitions. So I worked with a lot of startups that were turned into, like, the bigger Nutanix brand.

And then during COVID, I decided, okay, let's do something different in tech. You know how we all, uh, had this moment during COVID, like, what can we do for the better of the world? And I was like, let's try cyber security, because it's a, it's a different angle to tech. And I joined NoName Security, which, uh, as I said, is the company where I met my now co-founders, Shai and Avi. They founded the company back then. For about four years, was acquired by Akamai for $500 million, um, like two years ago. And that was the moment we got together and started thinking about Unframe, and, uh, yeah, we started building, raised, uh, two rounds, $50 million, and built the team since then.

Amit: How did you meet these guys? Like, so, so you were working together at NoName Security, that's where?

Larissa: Correct, absolutely.

Amit: And then so after the acquisition of that company, you guys decided, okay, let's, let's do something new. And, and then AI was around. Actually, we already left before the company was acquired, to be honest. But it was, like, because we felt this push or the pull, I guess, from the AI moment happening, and it was just the best moment in time to do it, you know? When, when you have such a huge, huge shift like AI, LLMs popping up, it's like, you now or never, uh, you need to take advantage of it. And so we, we did that, and it was perfect timing because just a few months later, the company was acquired.

Amit: How, I actually want to ask you some of the debates which are going on in the AI industry and tech world today. One of them being—so there are a lot of, um, you know, obviously every single day we see a lot of new demos on Twitter and LinkedIn, and it's obviously the vibe coding era as well. So a lot of people are creating products on the fly. And then, but you, for example, you are doing a lot of work with the enterprises. What do you think, like, is there a difference between vibe-coded products and things which look cool in a demo, versus the products which actually make an impact?

Larissa: Absolutely. Those two worlds could not be further apart. All of those have a real reason for existing, but you work a lot with financial services. You cannot imagine that the trading desk of a Wall Street bank is going to say, "Oh, cool, we have, like, some intern who just vibe-coded a new, uh, trading desk software. Let's hook it up to, like, the NASDAQ and let's get going." Like, that's completely impossible. So two worlds that just don't speak to each other.

I would say what's most important working with enterprises is like the reliability, the trust, the security aspect. Often it comes to integrating into legacy systems. We work with a lot of enterprises that have some built-in-house software from, like, 30 years ago or 15 years ago that's absolutely legacy, but they're fundamental to their business right now. A lot of it is manual, in people's heads right now. None of that is, like, something you can do with some natural language app generator and have the trust of senior leadership. That's just not reality at the moment. And often what people forget is, like, what about long-term maintenance and feature enhancements? Like, you can have a bunch of people in the company that create agents, that create apps with, with vibe coding techniques, but what happens when they move roles, when they leave the company? You just have, like, a bunch of rogue agents running around, like, operating your business? That's not the reality.

Amit: Right. Very well said. Another debate it will be great to know your views on is, um, so you know, like, uh, we call that horizontal AI versus vertical AI debate which is going on. Uh, I think the core of it is that foundational models are growing in terms of their capability almost exponentially. A lot of, uh, generic horizontal AI tools which are being created, they may be like a one-feature update to foundational models, and we have seen like hundreds of such companies, um, you know, going down, because for example, in video editing, uh, I have seen a bunch of companies, once Veo 3 and, you know, Sora and all that, uh, you know, got updated. Um, do you see like, um, do you see a world where startups could still continue building horizontal AI solutions? Is there, is there still a space for that, or all the startups should focus basically on vertical AI companies and go into some niches?

Larissa: It's an interesting one, honestly. Um, we've seen a lot of improvement on the LLM side, but is it really still going exponentially? Not sure. Like, in the, the last few months here, we've seen the LLMs kind of become a bit more of a commodity, I want to say. Um, even though there are updates, there are some really good players. It's not just the OpenAI anymore, right? Anthropic is doing great things. Google is doing great things. And I think no matter what, we are just in the first inning of, of what we're seeing in, like, AI meeting the enterprise right now. So what we always tell our customers is, like, you need to keep your options open. And especially with, with the importance of, like, business-critical use cases in the enterprise, don't just rely on a startup or a, um, like, even if it's not a startup, but like a more, um, experienced company, like a one that's been longer around. Like, don't put all your eggs in one basket. Like, keep your options open. If you want to change a model, choose a partner where that is possible. If you want to change the interface—like a lot of companies just have, like, a natural language, uh, kind of chat interface, the way that we're used to it from our personal use of ChatGPT. That is not the, the type of interface that you need for all of your business software. That's not natural. That doesn't feel comfortable for the user. So I would always tell everyone, like, be flexible, try things, um, and choose the partner that works for that. And there might be use cases where a very vertical play is the right thing to do. In many cases, I think, just speaking of our own experience, I think you can learn a lot as a startup founder if you have a bit more of a horizontal view and kind of get to know different, uh, industries, different teams, different departments, right?

Amit: And I think it's a question in general, but specifically in your case, because you're working across sectors and and doing so many things. When, when you're looking at, let's say, within financial services, you're working for, you know, some, somebody like, like a Realtor or or a small business lender, how do you develop that sort of understanding of the workflows and what kind of data exist and how do these processes happen? That a little bit of, I'm guessing, the domain expertise, uh, is required even to build any type of software for sort of solving that problem. So, how do you sort of cope up with that?

Larissa: Is it though? I know it's a very controversial thing, but we are winning RFPs very, very often against other companies that have strong domain expertise, but they have the AI capabilities as an add-on after the fact. So even though they might be knowing the real estate, the health insurance, the manufacturing business to a T, the performance often lags because they don't have, like, an AI-native brain, an AI-native, like, uh, backbone from the start. So I think you can do a lot there to kind of make up for domain expertise while you're collecting it. That being said, it's really important to have fun understanding the domain. You know what we do is, actually, it's not rocket science. We ask the customer, we observe the customer. If we are revolutionizing a specific workflow that they are doing manually right now, all we do is ask the question, like, hey, while you're doing it next time, can we be on a call? Can you share your screen, and we just observe you doing it? And so we understand how this looks like for you day-to-day, and then we build the software around their actual needs and pain points.

Amit: Right. Right. Okay. Very interesting. And, um, so far we have seen that, you know, a lot of AI products, you know, sometimes they will have like little issues here and there around hallucination. How have you overcome that problem?

Larissa: Yeah. The important thing is the tailoring. That's, like, the, the top one priority here. Because many tools that you see that promise you, "We can do enterprise search across all of your business systems," and you're just overloading it with, like, so much data that you have in the business that's not tailored to the specific use case anymore. So we really start with one use case. Like, what data do you need? And sometimes your integration source might be Salesforce, it might be your ERP, it might be SAP that is full of data, but you don't need all of that for that specific use case. So we really try to segment it so we don't overload the LLM and give it very clear instructions. We always ask it to reference the system of record again. So it's very easy for not only us but the user itself to see, where was this data generated from? And, uh, that gives you a lot of help already with, with citing itself. And depending on how our customers feel comfortable with AI, we add a lot of human-in-the-loop steps. For anyone who's just starting out, especially if it's, like, external-facing use cases we work on or anything like that, we suggest not having agents run rogue and, kind of, uh, take over the world in your company, but, like, let's put in a few steps of human-in-the-loop approvals, just as you get comfortable and, uh, and get your AI experience.

Amit: Since you are working with so many different, uh, companies across industries, um, I would love to ask you this question to see, you know, where we are in terms of adoption. So two parts to the question. One, which industries do you think are leading, uh, AI adoption? And then, if you could also mention that, okay, within this industry, based on, like, number of processes which are there and how many have been automated with AI, where are we today? Are we at like 2%, 10%? If you could, I know it's a, it's a little bit of a crystal ball, but if you can throw some light.

Larissa: No, I love it. I think the industries themselves—and we do a lot of, like, primary research. We look at, like, those numbers quite closely, um, just for our own interest and kind of, uh, outbounding efforts. Financial services and insurance, they do a lot of work. It's just not in production in many cases. They're trying a lot. That's what we see. The digital natives, there's a lot of things going on there, and they are using it in production, so tech itself, quite interestingly. But I think more so than not, I would say the industries where there are a few key players, that's the deciding factor. So think about in, um, like, Wall Street banking, in top-tier insurance in the US, in top-tier real estate services in the US. Like, those you have very clear set of, like, three to five key players that are very competitive with each other. And that's usually the, the areas where we see a lot of, uh, competition and moving very, very fast and trying out a lot of things that really make the difference in the AI-native world. But overall in adoption, I would say probably single digits. Even though they're trying, it's a lot of—I mean, you probably saw the MIT report at nauseam, but, uh, 95% of AI projects fail, right? So full production, single digits, in my opinion.

Amit: How, also in that study, it was very interesting to see that people just read the headline, but there was a lot of details inside it. So, for example, okay, maybe 95% projects fail, but then it said, uh, that in many enterprises, people were just trying to use ChatGPT or Claude to do something which was very specific and customized, and you needed something custom to be built for it, but they were just trying to use ChatGPT to do it. Obviously it will fail. And then they also said that in the cases where they actually used an external vendor specialized to do that, there was a 70% success rate. And that changes the perspective quite a bit. Is that, like, how you are also seeing it in your experience?

Larissa: Yeah, I think a very important factor around that is, like, even though you as a senior leader in an enterprise, you think, like, well, we cannot do AI. That's, that's sometimes we, like, our CEO would never approve it, you know, when they're early on in their journey. It's like, your employees will do something with AI if you want it or not. And in the worst case, they will put your company data in their personal ChatGPT subscription. Like, you can't, like, it is happening. It is happening. So you better give them some kind of solutions that they can use that you approve as a leader, that you feel good on the security aspect, that you feel good about from, like, a maintenance perspective, and that you have guardrails around. So you need to act. I would totally agree that that's what we're seeing.

Amit: Right. And, um, as a follow-up, within enterprises, there are different things, right, from marketing, sales, software development, operations. Where are you seeing the most amount of adoption today?

Larissa: I mean, we're not playing there, but I would say software development is definitely one of the big things. Like, I mean, I think everyone is using GitHub Copilot and Cursor, and, like, there's a huge reason for, for that being so successful. Uh, I would say that's, that's a key area, plus the users there are naturally curious in terms of new tech anyway. So I would say that is definitely the number one. I'm happy for our top use cases maybe not being in, like, the super sexy, shiny areas that are interesting. Like, I love the use cases where enterprise leaders tell us, like, "Wow, this is a problem and it has been for like 10 years, and we've been trying all kinds of different solutions and we haven't been able to. Wonder if you guys could do this." And I love those, because there's no solutions on the market for any of those use cases. It's very legacy to them, like very, very niche in a way that they wouldn't even know what to Google to find a solution. And that's really the stuff that I like most. This can be from operations, production, warehousing. Like, it can be really across the board.

Amit: How, given your experience in marketing and sales especially to enterprises. Now, if you look at it from an early-stage startup, usually people are very afraid of going to enterprises, and also for the right reasons, because it's very hard to get traction with enterprises when you're very small. But it looks like you guys had a lot of success, and you, you have been around only since like 2024, right? So can you tell us something to, like, young founders and early-stage companies, what they should do to sort of, you know, go in and penetrate the enterprises? What do you tell them? Like, how do you engage with them?

Larissa: Yeah, I mean, first of all, don't plan on sleeping in the first few years. It's a lot of hours. It's a lot of grinding and and pushing, and there's always setbacks. You know, uh, you just have to kind of pick yourself up the next day and start over again. But ultimately, what the enterprise cares about is the security aspect. You know, they need to feel comfortable that you know what you're doing. And it really helped us having a cyber security background and having a track record there. We built all of our security controls, all of our security certifications immediately when starting a company. That was, like, I think after the incorporation and and the funding papers, that was kind of the next huge stack of paperwork that we did, because we just knew, if we don't have those capabilities, it will be really hard for an enterprise to trust us. I mean, obviously, we, we also built, uh, our product in a way that there's no cost involved until they're actually happy. So that kind of lowers a barrier. But I wouldn't recommend that to everyone, because it's a really hard thing to do, and most companies are not able to do that. So don't, don't listen to this advice here. But yeah, I mean, I I understand it's scary for many. For us, it was kind of what we had always done. So I would say it felt less scary. I honestly, if if my background just thinking, like, I think doing performance marketing and like having to gain customers, like SMB and and individual people through social media performance, I was like that would be very scary for me. Like, enterprise is, uh, what we've always done and know how to do.

Amit: Right. Just last two questions. Uh, one is, is there any particular case study that you can actually talk about? That one or two case studies very quickly that, you know, for this particular customer, this kind of problem, we solved this.

Larissa: Yeah. Um, so I can just tell you something I just discussed on a customer call this morning. So it's, it's top of mind. We work with a really large, um, and very reputable newspaper. So media company, basically. They still do print newspaper in this day and age. And with everything that's been going on, you know, you used to have a huge proofreading team that went through all of it to check quality, to check spelling. But now there's so many articles across the print, across web apps, all of the stuff that that newspapers need to stay competitive. And they just don't have the bandwidth anymore to proofread everything. And that's an area where AI was really able to help them very, very quickly. Just think about it as, like, a co-pilot to, to the actual editors and then the writers itself. Currently, in order to hit the quality, they needed about two to three years of onboarding time for a new proofreader. But with our solutions now, they were able to reduce that time to immediate, basically no onboarding time. I mean, give or take a couple of days here and there to understand, like, your bearings of working in a new company. But, uh, that was a huge, huge, um, efficiency gain. As well as the time that it took them from proofreading an article. We were able to lower that to like 70%, like, reduction. And that made a huge impact. And I think it's a really cool story, because no one would think the media industry, who's already scared of AI, and then traditional newspaper, is a good starting point for for AI transformation.

Amit: Absolutely. That's a that's a very good example. Um, my last question is that, apart from your company in the AI space, is there another company which you admire or you think, like, you know, they're doing great?

Larissa: Interesting. Honestly, I feel like I see some of them every single day pop up. There's so much cool stuff going on. Do I think that from a business perspective they are all going to survive? Probably not, because a lot of it, um, as we have seen, as you mentioned, with LLMs improving, um, that they're, they just, like, don't really have a reason of existing, unfortunately. But there's so many angles. You can see that there's some that I'm, like, they're close to my home, and I think they're, they're doing something really cool. They pop up everywhere in, like, customer support. And then there's others in, like, the legal space completely taking over. Not just, like, building software, but like those full-stack AI companies that that pop up. I'm not sure if you've looked at it, um, a lot, but they don't just launch software. They are actually building a whole new legal firm, a whole new, uh, real estate company, like everything AI-native from scratch. I think we'll see a lot more there, and I think it's, uh, it's really fascinating.

Amit: Larissa, thank you so much. This was wonderful talking to you. Very, very insightful. I'm sure our audience will learn a lot. Thank you so much for your time.

Larissa: Thank you. Uh, love this energy in our conversation.