He raised $75M—but kept his team to 21 people & pays everyone the same salary. | Cos Nicolaescu, Co-Founder of Accrual

Cosmin Nicolaescu (00:00:00) : The reason why you're here from a financial perspective should be the equity, not the cash. Otherwise, don't join an early stage company. It's not worth it. If you don't think that early stage company will 10x, 50x, 100x, whatever. Something very, very substantial, you should not be at an early stage company. Go work at a big company. You'll have a much better quality of life. It's very easy to demo stuff in AI. I can build a demo in the next few hours on almost anything and it will impress people. But to actually build a product that can do it at scale and reliability necessary for production, and accuracy. Especially in such a regulated industry, is exceptionally, exceptionally hard. Everybody at Accrual has the same salary. Pablo Srugo (00:00:35) : That's wild. Dude, I've never heard of that before. That's crazy. Previous Guests (00:00:41) : 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 (00:00:53) : 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. Cos, welcome to the show, man. Cosmin Nicolaescu (00:00:53) : Thanks for having me, Pablo. Pablo Srugo (00:00:53) : Excited to have you, dude. You just raised a massive Series A, $75 million Series A. Dude, the Series A's are just getting bigger and bigger, man. It used to be $5 million in an A. Cosmin Nicolaescu (00:01:17) : Nowadays, it's actually. I guess, small if it's not hundreds of millions. Pablo Srugo (00:01:21) : It's all relative, right? Because I had a founder tell me, yeah, I got the $10 million ARR but it took me like three years, I think he said. It's so slow. I'm like, dude, come on, man. Cosmin Nicolaescu (00:01:30) : Nowadays, it's like if your first month is not $10 million and $100 million in a year, you're way behind. Pablo Srugo (00:01:35) : It is, I mean, it's great when you have it and when you're not the top one percentile, you're like, oh my God. What am I even doing with my life? Cosmin Nicolaescu (00:01:42) : Yes, I think the fundraising story is very split now between the types of businesses and founders, and everybody else. Pablo Srugo (00:01:49) : We'll get to all of that, but let's start where we always start and then we'll deconstruct. I'll talk about what you're doing and how you got there, and all that. But to start with the thing that matters most to this show, right? The product market fit moment. Tell me, for you, when did you feel you'd found true product market fit? Cosmin Nicolaescu (00:02:05) : I would say, so we started the company in October '24, and we had a lot of ideas. And a lot of hypotheses of what we can build, but we had no idea how much of it is real. What technology can actually do in this space and how deep you can go, and how much moat you can actually build as a business versus just swapping some low hanging fruit, and then being done. I would say the last fall, was supposed to be about a year in, is when it felt like, okay, this actually works and clicks, and all the moving pieces. Because prior to Accrual, most solutions that I saw in the accounting space were very point solutions. Where you're taking some type of documents and extracting them, or taking a client portal and showing it to your customers or whatnot. So very, very niche things. Which don't really challenge product market fit. You're not creating a category, you're not doing something unique, you're just hopefully doing something incrementally better. Accrual was started with the idea of building a platform for these accounting firms, which means you have to have a lot of these components and parts of the workflow that all click together. And you have to make sure that AI actually can learn things, because it's very easy to demo stuff in AI. I can build a demo in the next few hours on almost anything and it will impress people. But to actually build a product that can do it at scale and reliability necessary for production and accuracy. Especially in such a regulated industry, is exceptionally hard and so for us, last October is when we did our first set of returns. Where we basically took very complex documents from firms like Armanino, Creative Planning, some of the largest, most complex individuals in the United States. We ran them through the entire system. We had people actually using the platform and the product, not just calling things through APIs and processing them, and interacting with the platform. Understanding how the agent worked, what it did, and the amount of savings that people saw in terms of impact was just staggering. A month before, we had no idea if it actually could work at that level. It was much more piecemeal and isolated than when we did it. So that was, I think, the first time I was like, okay, I think we have something here that actually could work and scale. And that put us into this tax season, and it's all from there. Pablo Srugo (00:04:22) : Would you say, there's a lot of people who have a product that some people use, and even a product that some people get value from. But they don't necessarily get to product market fit, right? Like, true pull, hair on fire solution, whatever. What about what was happening then felt like real product market fit to you? Cosmin Nicolaescu (00:04:41) : I think the signal that we look for is the same signal we had early on at Stripe from folks that were using us, Which is, what's the excitement level with the customers? Because there's a lot of times where people are like, yeah, this sounds great but quite skeptical. Then they're kind of like, oh wow, there's something there. I think that's the next level where you're solving a problem, this is quite plausible. Now we're at a stage where, and this started happening in the fall. Where not only do they have the level of excitement, but it materializes with commercials. Because I think that's the kind of rule that we learned early in our careers, which is it's one thing for people to get excited about a demo and trying something out but until people pay for it, you don't know the true value and once people are willing to commit to complex, large contracts. That's when you clearly have something. Otherwise, people wouldn't do that. But the biggest reward that I think we see in terms of personal satisfaction and validation of product market fit is, when people are like, well, can you also do X? So now your clients are pulling you into their business to help solve more problems. That I think is the highest level they can say. Which is, you've not only solved this problem and I'm willing to work with you, and roll out your product. But also I'm now eager to get you to help me solve other similar problems across my business and that purely comes from internal confidence that people have. Otherwise, obviously, they would never do that. Pablo Srugo (00:06:12) : How do you tease that apart from. Because I agree there's something interesting where you said, there's that failure mode, which we talk a lot about. About one more feature, right? Where customers will come to you and say, if you build this, then I'll buy. What you're saying is similar, but different. How would you tease those two things apart? Cosmin Nicolaescu (00:06:29) : Typically, people either gravitate towards areas that look similar in terms of problem solving. It's like, hey, what you're doing here, if you tweak it a little bit, you can apply it here, and we certainly had that. So as an example, we started with individual returns and then people are like, oh, could you help in other types of tax returns? Very natural evolution of the product. The second type of asks are completely different, but big pain points and so we've had customers. Actually even in this past week, they're like, hey, engagement letters are a pain in the ass. We haven't found something good that helps us with that. There's a lot of these products and they're all very limited. Could you actually build something there? Nothing that we've shown them or built that has that, but there's a confidence of the quality of work that you've shown and the impact in this area is so high that we believe we could apply that same level of talent in another area that is very painful for us. We saw the same thing with audit, where we've had the partner from audit in different firms be like, hey, I've heard all the wonderful things you guys are doing on the tech side. How can I start using you guys on the other side? Because we have similar problems with existing tools and processing workflows. And if you guys could do something similar on the audit side, that would be great. Now doing the same thing on CAS. So you can see that FOMO that trickles within your organization. Pablo Srugo (00:07:50) : And to be clear, these are not potential customers who are saying, if you build this, then I'll buy. These are actual customers for the most part who are saying, I like your product. It's great. I'm buying. Can you also do this other thing? Can I also pay you more for this other thing? Cosmin Nicolaescu (00:08:02) : Correct, yes. You're already solving this problem and I'm rolling this out. Can you also do more? Pablo Srugo (00:08:07) : Yes, which is a big difference. The other thing I wanted to dive in on is, the value prop. So now we'll talk a little bit more about what your product actually does but that's the key ingredient of product market fit, right? It's not just they're using it, they get value. It's what value are you really delivering to your customers? Cosmin Nicolaescu (00:08:25) : So for us, we look at it from two perspectives. One is, how are we helping the accountants directly? And most of it is productivity, quality of work, enjoyment. Are you doing things more productively and are the things that you're doing the ones that are more interesting in your job versus boring, annoying, mundane. Whatever you want to call it? I think that's part of it. The other part is the client experience. A lot of the firms, and that was one of the things that attracted us to this space, is accounting firms have worked with their clients for a very long time. They deeply value those relationships, and that's probably the most valuable thing in the business outside of the people that they have in the firm. And they're very focused on like, I don't know how I can provide a better client experience for my clients. And if I'm able to do that, then I believe I will be more valuable to them. It will be even stickier because I think people are realizing that over time, technology will solve more and more problems that are mechanical in nature in this space, like filling out things, extracting information, some set of workflows. But the client relationship is the most valuable thing and that's where the value creation exists, and the new set of services. And so, a lot of them are trying to optimize for that as well. And that is a big differentiation in the pitch of having an end to end product that helps both the accountants and their clients feel better. Pablo Srugo (00:09:46) : And maybe just tell me a bit about the before and after. Especially when you were talking about this product market fit moment specifically. What you delivered at that point? What were they doing before? And then because of your product, what did that enable them to have their day be like? Cosmin Nicolaescu (00:09:56) : So before, let's just talk about 1040 individual taxes. Because that is what we just finished our first tax season at scale with. Before Accrual, the best accounting firms in this country and the most innovative ones would have anywhere from three to four tools to six, seven, eight tools. That's kind of the range and so they would have a tool that was their client portal. They would have a tool that would keep track of the intake forms that customers have to answer and upload documents. They would have a tool to store documents. They would have a tool to extract information from said documents. They would have a tool to review things. They would have a tool to have the client review stuff. They would have a tool to actually sign stuff. So each step basically had one or two tools, and part of it was the ecosystem that has been created. Part of it is the evolution over time. Like, oh, you were using this tool and then we started using this other tool, and now we use both tools. Because that's the general saying. How do you solve a problem? You add a tool and now you have three problems. And so, that was the landscape. And firms using Accrual, they basically get excited about the impact that we can have, but they also look at it from a pure vendor management standpoint. It's like, oh, I can collapse many of these tools into a single tool and it's both useful for me as the firm where I have less tools to manage, less trainings I have to do for people, et cetera. But also for my clients, to give you an example. There's two places where clients interact with their accountant in the tax workflow. One is gathering information. The other one is reviewing the tax return and signing off. You would imagine that those are the same tool, because that's what the client is using and it's not. You literally have, as a client, at the very minimum, two different portals in various forms that you had to interact with for different stages. That's kind of crazy that that's the world that we used to live in and then in between, you had a variety of these tools. And not only were the tools not that helpful in terms of productivity, but they required so much manual work from admin staff to go in and out of those tools, import and export and transform data, put it in Excel, re-save it. It just feels like a very leaky bucket. Pablo Srugo (00:12:06) : The problem of having too many tools, it's classic, right? Especially with B2B SaaS, that was the whole thing. You have all these point solutions, blah, blah, blah, and trying to build an all in one was always the gold standard. That's a classic pitch, the all in one solution. I'm seeing with AI, it's coming back, but it seems to be working in a way that it wasn't working before. My question though specifically is, how do you make that work? And if you think about that first customer or maybe one of those first customers as an example, how do you tell the customer, trust me, you get rid of these six tools and you'll do it all on mine. It's going to be great. It's going to be fine, right? It all sounds great in theory, but how do you get them there? Cosmin Nicolaescu (00:12:41) : I mean, obviously, higher risk, higher reward in that we have to build a lot more. It'd be much easier for me to build any one of those tool replacements with higher quality, but the gains are much more incremental. So you don't get that order of magnitude productivity gains. I think there's two things that have resonated. The most important one is we've started in the beginning and we still do with a vision of what we're building, and obviously that makes sense in our first year. Because the vision is the only thing that we had. We didn't actually have a product yet, right? We hadn't built all these things. But even now, people are buying the vision of Accrual, helping them be the operating system across their firm without having that capability. Because we don't have all these different products and service lines in place. But that's landing that message. It's, hey, here's one, why this world will look better. Because you've been used to a world in which you buy multiple tools. I was at a conference earlier this year, and one of the partners on stage was advocating for using the best tool in the best scenario. So combine all these tools and I went in afterwards and I was like, hey, actually, I don't think that's going to work anymore. I think that's actually counter to that and I want to explain why. And that's the reason number two. But that's not a common pitch. So first, you have to paint the picture of what that might look like and then you have to convince them that you can actually build it. Because to your point, you haven't, and so I'm taking a risk in doing that. And that's why having some level of credibility and showing iteration very quickly, I can build that credibility when you don't have a product in place. Pablo Srugo (00:14:15) : My feeling, I'm curious on the pre Gen AI, post Gen AI, and how that affects this exact pitch of this all in one solution. One of my, let's say, educated guesses or whatever is that there was always value in the all in one. Because, obvious reasons, everything you just said. But with AI, it does feel like there's an extra level of value of being able to have the context go across. Do you think that's true? Cosmin Nicolaescu (00:14:34) : A hundred percent and this is exactly what I told that partner. If I look at software engineering, which arguably is the most advanced field of using AI right now, and even then. It's like you read about Silicon Valley tech startups and the biggest tech companies in the world using agents for coding. The majority of the companies in the world building code are not, and the reason why it works so well is exactly that, which is I have all the context in the world in my repository. That is my world. I don't need anything else because that's how code works and I'm able to iterate through, structure things, test them, have that feedback cycle be as close as possible, and know the entire universe. I think that is very different where I have an agent that extracts some information from a document and then gives you the result for that, and then another process that takes those, and wants to use it. But how many of those documents were related? And how much do you know about your client as a result of that? And what do you do with nuanced situations if you just get some raw numbers that are OCRed? You just lose all that context and all of a sudden, the capability of the model is diminished significantly from what it could do. And so I very much agree in exactly what I told that partner. In this world, I think having that context is not only useful, but critical. I think without it, you will not get the gains of the promise of AI that you see in other industries without doing that. Pablo Srugo (00:15:59) : It's the comeback of the all in one and it also raises the bar for choosing correctly. Because if you had a one point solution and you picked wrong, you just swap it. But if you go with the wrong partner, three years out, you're going to be behind a lot of other people. Cosmin Nicolaescu (00:16:13) : And you see that a lot. Another thing that we were looking at initially, to your product market fit point, was how much are we building a ChatGPT wrapper around these models versus going really deep there. And obviously, we went down to build workflows really deep and understand that domain. But a good litmus test to see whether you're building this all in one platform that has very high stakes but potentially high rewards is how many of them can you try out at once? Not as a pilot, you can pilot multiple tools, and I think that's fine and encouraged, actually. But how many can you actually deploy live? So if I give an example, let's take a research tool, or even, I want to give everybody ChatGPT. I don't know if ChatGPT is better or Claude is better or Copilot is better. I'm just going to give license to all three of them and people can use any of them, or different people get access to different ones and see which one ends up being productive. Very low cost of change management, and thus very low cost of deploying. If you're using ChatGPT and tomorrow I tell you you're using Claude, you're going to switch pretty easily between those as a just consumer. Which is why we've seen people switch between these tools as an individual. Very different when you go very deep on these workflows and you're now rewriting the standard operating procedure, and you're sending stuff to your clients. You have various people in different roles throughout the workflow using your product all in one. I think a single use scenario, the single player scenario where I'm doing some work with some tools is very different when I'm working as part of a collection of people throughout the workflow and all of us use the same tool. And I think the latter, is way harder and you're taking much more of a risk as a firm in deploying that. Because the cost of switching is indeed much higher, which is why all this existing software is actually so sticky despite the quality of the software. Pablo Srugo (00:18:07) : So let's go back and go through the story. Now that we understand the value and the product market fit moment. You were a CTO at Brex and, walk me through your decision to leave and start Accrual. What's the origin story here? Cosmin Nicolaescu (00:18:19) : Yeah, so quick highlight of my career. Originally grew up in Romania, moved here for college to study computer science, and then started my career at Microsoft. So worked at super, super large companies, worked on Azure, Office 365, all the cloud services. Then I was like, I want to do something much smaller. I moved to San Francisco, joined Stripe. I think it was about a hundred people when I joined. Got to build all the payment rails, got to work with my now cofounder very, very closely. We were on the same team together. Then joined Brex to repeat that, but at a different stage. I can run a function now and join even earlier. I joined Brex, I think it was thirty, forty people. I was there for about six years, building the function. After all that time, I've always told people, unless you're the founder of a company, you shouldn't be at the same place forever. I very much encourage people not to jump ship every one to two years, not just because it looks bad, but because to actually see, especially as a software engineer, to see the impact of your work takes time. It's one thing to build something new. It's a different thing to scale it. It's a different thing to iterate on that. Pablo Srugo (00:19:23) : It's like the Tobi tour of duty idea. Like come here, do your five years, make some happen, move on. Cosmin Nicolaescu (00:19:29) : Correct, so for me, three to five years, people staying in that range feels quite good. If you stay longer, there's nothing wrong with it. Obviously, people have had immense careers at these companies that are rocket ships. If you're on a rocket ship, hang on and learn as much as you can. Pablo Srugo (00:19:43) : Yeah, that's right. Cosmin Nicolaescu (00:19:44) : If the company's not good, I get it. But if you have that pattern, there's probably something not great there and so for me, six years. Obviously, I was in that kind of range. I look at it mostly from, where do I think I provide the most value and where do I get most excitement? Having gone to smaller and smaller, and smaller. I just internalized that I enjoy building and being closer to what gets built, than running large organizations and scaling companies. And my path at Brex, if I would have stayed, would have been to be the CTO of a public company or now that it got acquired, be part of a larger organization. That was kind of the, I am becoming an executive at large scale and that wasn't super attractive to me. Again, I enjoy building. I don't enjoy kind of coordinating things across the org and aligning and reviewing as much. Very critical, and I think there's people who are exceptional at that, just not where I get energy from that. And so, that was kind of my plan of okay, I probably should hand this off to someone else. And we had someone exceptional on my team who I've known for many, many years that took the baton and has been doing phenomenal there. So also a good growth opportunity for someone internal to transition. Then there's a version of Arcave, now what do you do? And I've just like, I take the approach on building the harder parts of the product first, the same thing is true in my career. Which is if I'm not challenged and I'm not learning, I get bored very quickly. And so, the idea of going and repeating and being a CTO again. And going through another wave of something similar even after a different industry, different founders, wasn't that appealing to me. I kind of knew what that looked like and I didn't feel like I would learn as much. Going back to a big company, again, not exciting. I would have stayed at Brex, otherwise, and so for me, being a founder felt like the next challenge that I've never done before. And the only thing that I knew for sure leaving Brex after taking some time off is that Sid and I want to build something together. I would have not done this alone. People that do this alone, mad respect for them. I don't know how they do it. It was very much we're going to build something together. We'll figure out what it is. We'll do things our way and try to see if building a company in the shape and kind of culture that resonates with us can be successful. And get to learn what it's like to be in a different role. I've been an engineer, I've been a manager, I've been an executive. Now I can be a founder and take the beating from that. Pablo Srugo (00:22:05) : I envy the founders who just have this problem happen to them. You know, like, Tobi from Shopify story, and they go out and they solve it. And it becomes this massive business because that feels, I know I'm wrong saying this, but it feels like easy mode. At least at inception. It's kind of you just get pulled into something, you do it, and then executing is obviously hard but starting is relatively easy. Because it's natural, it's organic. What you did, I consider hard mode, which is, I don't know what I want to build. I know that I want to work with this person, but I have no idea what to build and especially, you've been and you've seen greatness, you've seen the successes. It's almost harder in a sense, when you're right out of school, you're just whatever, let's just build some shit, who cares? Where you have the bar of, this is with Stripe, this is Brex, it needs to be at least at that level. So I'd love to go as specifically as you can through that period, I guess it would be like early '24, where you and your cofounder are trying to figure out, okay, cool, what do we build? Cosmin Nicolaescu (00:22:59) : Yeah, so we spent, I would say, a hardcore three to four months. Loosely five to six months, brainstorming, iterating through ideas. It's interesting that you say easy mode or hard mode. I always look at it as there's some founders who know their mission on Earth is to cure cancer or solve whatever problem or industry and in those cases, it's relatively exact. This is what I'm doing. Infamously, Zuck, when he had the offer to get acquired by Microsoft for a billion dollars. Basically was if I sell this company, I would go and build another social media company and already have one that I quite like. So why would I go do that? Elon, similarly, wants to make humans a multi-planetary species. Everything kind of goes towards that and you have a lot of the founders that have that, like Brian Armstrong with crypto, Patrick and John at Stripe. Very clear, this is my life's work. Neither Sid nor, I had that in particular. It's true even on a personal level. I donated to charities of random things because I empathize with a lot of those things. I want puppies to have a better life in shelters. I want kids with autism to have a better life. I want older people to have enjoyed the last years of their life. I want to solve any disease. Pablo Srugo (00:24:18) : I want to fix all the problems, all of them. Cosmin Nicolaescu (00:24:20) : I don't get special pleasure from one versus the other and so for us, the main thing was impact. Anything that we do has to have very high impact. Because for me, it's like learning and impact are the two things in my life that have motivated me the most. Whenever I've been bored, I kind of spiral out and then don't do well, and lose focus. And if what I'm doing doesn't have impact, that it's kind of like why am I? Being a founder, working at an early stage company, even as an employee, it's just like so much work that if you're not having the impact, why are you doing this? And so, everything that we kind of shortlisted was industry wide impact, transformational and not only that but the transformation of that industry should result in something better for a much wider group of people. So for us, the motivation for accounting was seeing what Stripe and then subsequently for Brex operated at, when finance and accounting became very strategic. At Stripe, we became very strategic when Will Gabriel became CFO. At Brex, it was from the beginning. One of our first hires was our CFO. Both cases, very business oriented. Finance and accounting was in the room for most important decisions, and that results in generally better decisions. Same thing on individuals. Every wealthy American that you know has a family office, has accountants that help them make better decisions and optimize for the long run. And so, how can we actually bring that to every business? So accounting is not just the back office kind of function that just ticks some boxes. How do we help every individual have better financial planning and financial decisions, and education from their accountant? And more people being able to have access to that because it's such a small percentage of folks that have the ability to do that. And so, that was kind of the motivation of why we want to go in this industry in terms of impact and how will that impact actually benefit society more broadly? Pablo Srugo (00:26:11) : And how do you get there? I'm curious, you guys sitting around and you're kind of listing big industries and saying, well, that one's already kind of solved or, Harvey's already going after legal. So no, or are you talking to accountants? You know what I mean? What's the process? Cosmin Nicolaescu (00:26:24) : So we started with a list of what do we not want to do? What is it that we're just not particularly passionate about? So to give you a couple of examples, neither of us are especially fanatic about crypto. That's not to say we're bearish on crypto, but to be successful in crypto you need to be fanatic about crypto. My buddy Zach had a product at Bricks, became founder at Bridge, now at Stripe. Pablo Srugo (00:26:43) : Yeah, he was on the show, by the way. Crazy story, man. Cosmin Nicolaescu (00:26:47) : He was just obsessed about that. Pablo Srugo (00:26:49) : Right Cosmin Nicolaescu (00:26:50) : Every person I know who does well in crypto believes that is the future very, very deeply and wants to make it happen. So I was okay, we're not super passionate about that. We shouldn't do anything that foundational labs are doing. We're not research people. We're very much kind of applied AI people and product people. And if we're going to do AI model related stuff, We should just go work at OpenAI, Anthropic. They have smart people, they have funding. We know the founders at both places. What are we going to do better than them? Next level was things that are adjacent to the models that the model companies will likely end up doing, even if it takes them slightly longer. Probably not a good business decision, even if it's technically interesting, and you can do that. And then finally, consumer products. My entire career, both for me and Sid, we've worked on B2B software. Consumer growth looks very different. I think there's people who are exceptional that I don't think I am. And so maybe I can learn it, but I should probably lean in on the things that we're good at. The things that we're very good at are gnarly problems with complex workflows, lots of moving pieces that all are intertwined and have to click well together, super high reliability, super high accuracy, super high security. Those are the things that we've spent most of our time and get a lot of energy from. Even within Stripe, we spent a lot of time with the finance and accounting teams building the ledgers. And figuring out how to close our books more efficiently. We got energy from that, not everybody at Stripe did, though people were interested in other parts of the product and that's great. Different people gravitate towards different things and so once we had, here's what we don't want to do, and here's the rough shape of what we're generally maybe have an edge on. Let's look at where we can apply that. We looked at a wide variety of professional services. We looked at software, hardware, security, developer tools, HR tools, very, very broad. And so from there, you start looking at positives and negatives. So you mentioned Harvey with legal. That wasn't the reason why we didn't go deeper on the legal field. It was actually much simpler than that, which is we end up having much more enjoyment spending time with accountants than with lawyers. I keep hearing from people all over that there's far fewer accountants in this country than we need. I've never heard one person say we don't have enough lawyers in this country. Nothing against them in any means. I think they're doing amazing work and very critical. But if I have to choose between the two, that was a very clear tiebreaker for us. When we looked at hardware versus software, very interesting to work on hardware, especially nowadays, everybody's basically oh, software has been solved. Work on hardware, we would have learned a ton. We realized that the things that we're, again, very good at are the software parts of the hardware. Even then, it's a very different type of software that you have to write for hardware and so would it be interesting for us to learn? Yes, but is that the winning team? If I go to clients or investors and go, hey, bet on us because of our backgrounds in payments and help us build rockets. I would probably be I don't know, do you guys have someone who knows anything about building rockets kind of thing? And so then it goes back to leaning in on the things you should have some experience. Because one thing that is different for us is we're kind of like experienced founders in terms of like we're not just out of school and just becoming founders. Like Patrick and John and Pedro, we're like that. Dropped out of school, started a company, it's unbelievable. You're kind of that type of founder which is naive but incredibly optimistic and just get stuff done. Or you have founders who have worked at other companies, are much more pragmatic, and then you kind of bring that experience to make up for the lack of naivety. There's something very magical about I have no idea because I'm a 20 year old and I'm just going to go through the wall and then if I'm right, it works out. But having that confidence is something magical and over time, you kind of lose it because you're like much more pragmatic and practical. It's well, I've seen this game many times and more often than not, it fails. Pablo Srugo (00:30:51) : We have tens of thousands of people who have followed the show. Are you one of those people? You want to be a part of the group. You want to be a part of those tens of thousands of followers. So hit the follow button. I think that's right. I also think it matters how you're coming up with, going back to the original question, how you're coming up with the idea in the first place. Take Zuck as the crazy example, the super outlier one, he didn't set out to build Facebook is just like doing stuff. That happens, boom, and you come, oh my God, okay, I'm gonna dive into this, and he happens to be like a maniac and just crazy execution, so he does it, right? But if you're gonna go so top down like the way you're doing it. It pays to have that background, that experience, and think about team market fit and these sort of things, because, it's just, it's a different ballgame. It's a different ballgame altogether. Cosmin Nicolaescu (00:31:34) : If you're much more pragmatic and much more systemic. Then you have to kind of bring that across the stock. I think that's absolutely right. Pablo Srugo (00:31:41) : And so how do you go from, so that's actually super helpful to understand how you land into this space of accounting. How do you go from the space to the idea? Do you then do classic customer discovery? How do you figure out what this first product is going to be? *Cosmin Nicolaescu (00:31:55) : The lesson that we learned early at Stripe was the best way to build stuff, and ironically, the easiest, is find a customer that's willing to trust you with a hard problem and if you solve that problem, they'll give you more problems. It's as simple as that. And it sounds very, very simple. Obviously, there's a lot of work into that. But if you're able to do that, you can repeat that motion over and over and over again and build a very, very successful business. And so we went, talked to accountants, looked at what they do, shadowed a bunch of them, talked to owners. We understood kind of the broader picture issues that they have with the workforce and everything. But in terms of like which space to start off with was very much driven by a lot of customers and these large firms asking us about it. So our first two customers were H&R Block and Armanino. And very fortunate to be able to work with those kind of tier of companies. Armanino, top 20 accounting firm. some of the wealthiest individuals in the world and a bunch of tech businesses that work with them. H&R Block, huge amount of volume. Basically, about 20% of Americans do their taxes on H&R Block, to give you a sense. Very different shape. The domain complexity for H&R Block is very different. The people that do their taxes, generally much simpler use cases than the wealthy, high net worth individuals. But huge amount of scale, huge amount of regulatory complexity as a public company. very complex change management, very high security bar required. So very good forcing function to build a very strong foundation in this space that without that, we just cannot land that customer. And then Armadino, very high complex domain for the type of returns that they do and the kind of companies that work with them naturally. And the common thread between both of them is they both serve individuals, very different ends of the spectrum in terms of complexity. And in the cases of these top accounting firms, for them, individuals tend to have a wide range of this complexity to the point that there's many individuals that they have to serve that they wouldn't unless they would, that's the only way to kind of get a broader business. So I will do your returns for your business, but now I have to d o the returns not just for you, but your family and your friends and your employees and so on and so forth. Do I make a lot of money doing it for your employees that might have a simpler return? No, but I have to in order to get the more complex, higher engagement fee return. And so accounting firms, the largest ones, always look for opportunities to kind of improve both performance and efficiency of these, as well as the client experience to make it faster and better for them to not spend as much time on that. And obviously H&R Block, that's all they basically do. That was the first product and it was very much one that allowed us to build a foundational box. We were not strongly opinionated on which one. If we would have gotten pulled into doing tax returns for businesses, we would have done tax returns for businesses. of what we're doing now, any of them would have allowed us to build a lot of these foundational blocks. And that was most important. Can we find the right customers and design partners with the right ingredients for us to build these blocks that we can stack over time? Pablo Srugo (00:34:57) : So to move to the deep tactical segments here, and one of them is selling to enterprise and specifically converting from pilot to full customer. You have a hundred percent pilot to production conversion rate. And maybe that ties into this, because that was going to be my next question. Anyways, like H&R Block, Armanino, For those to be your first few customers, obviously you're drawing on your background and your credibility as someone who is a CTO of a very well-known company at Brax and so on. But it's a huge leap still for them to be yeah, sure, let's work together. How do you get in the room? How do you structure that and bring it down to a pilot that you can clearly deliver on with the resources that you have? Tell me all that and ultimately that leads into this getting pilot to production conversion. Cosmin Nicolaescu (00:35:40) : I would say the first thing that Sid and I felt very strongly about is being truthful to our customers. I think this is true in any industry, but especially in accounting. We've seen so much vaporware and so much promise and under-delivering on those promises. that erodes so much trust. And so from the beginning, we're we're going to be very transparent about what we do, what we don't do, what we will do, what we won't do. Ideally, we find people that are willing to work together versus be surprised. Because we've seen that over and over again, where you tell customers to do it, they get excited, and then you have to deliver on that. And if you can't, you're in a really bad place. Just starting with that attitude brings a lot of credibility. I do think our backgrounds probably helped in the beginning when we had nothing else, and compounded with this, it helped. The second thing is iterating extremely quickly. Sid and I, when it was just the two of us building, we would basically iterate on an hourly to daily basis. Like we met with you today. By Monday, we would show you some stuff. By Wednesday, Tuesday, Thursday, every day, there's progress being made and you get to see something that you're just not used to. So people see that and they're able to project it forward and be well, if every couple of days, this is the kind of progress these folks make, if I go a month from now, they will be so much further ahead. Here's what we'll do. We move quickly. We've done the thing. Now let's go to the next thing. Rinse and repeat over and over and over again. Builds that credibility to kind of take that leap of faith. And the last thing I would say is, because we're both engineers, we're extremely structured in our thinking. It's just how our brains operate. And so we take the same approach to running pilots. We're okay, we're going to run in this way with these people. We're going to set this up. Here's the different Pablo Srugo (00:37:28) : Yeah, can you walk me through an example of that? Because I think that's where the rubber meets the road. How do you boil down to something? How do you structure it and make it so that it's clear, it's winnable, it's quick? Cosmin Nicolaescu (00:37:39) : Yeah, so we obviously demo stuff. That already stands out because at least in accounting apparently there's still product out there that you don't get to even demo it properly, which is kind of crazy to me. But we run demos. We typically run several of those with different people depending on the firm. Then we said, okay, to get the right signal on a pilot, you need to find the right balance between too little and too much. So we've seen firms that are very hands-off, very little usage. And so you go through the pilot very quickly, but at the end of the pilot, they don't have high conviction because they're well, yeah, it was good, but we only did like a few returns, for example, so maybe we should do more. And now you've reset the clock on that. And so now it's great, now I'm running a second pilot. The other extreme is we're going to run hundreds of returns with dozens of people on our firm involved. Also very diluted because it's very hard to keep engagement high across a large number of people when you go through a lot of returns. If you run a handful of returns or a certain complexity or a certain shape, you're not going to learn much more from running dozens of those. You'll get the signal pretty quickly if they're good or not. But having a lot of data makes it very hard to go in depth. So we go, when we run a pilot, through returns of different complexity with different folks on different roles. And then we literally go return by return and give you a side-by-side comparison between your return that you finalized and our draft. If they're the same, great, they're the same. Not much to spend time on. if they're not the same, here's why they're not the same, and what led to that. And maybe it's on us, maybe we didn't have document or context, whatever it is, but it's very in-depth. And that helps bring credibility because now you have enough volume and enough variety in kind of the data that was tested. And you've gone really deep on all of them to get a good sense of will this product work, not work, for what does it work, what does it do, not do, and much more confident to move forward for them. And I think the third one is, again, no smoke and mirrors. I've heard of pilots where you get something and they have AI, but then they have humans that go and correct things that AI got wrong and then they give you the final thing. And it's well, that's not really the AI part. You should be upfront about that if you're doing that. I've seen others that will go through the same return multiple times and process it. Because again, many of these systems are not deterministic and so you'll get different results. And then they take the best out of all these combinations and give you a final thing. I was no, that's not how real world looks like. You should be upfront about it and just run it like your customers would because when they'll run it, they'll just see these failures and be disappointed. Pablo Srugo (00:40:21) : And what about the flip side, which is on the customer side. Keeping them close, motivated, on time and getting you know? Because the flip side of the failure mode is you actually do everything you're supposed to do on your end. But maybe you don't get enough, attention from them, in your case, returns from them, or maybe you do and then now you're waiting a month for a decision. How do you kind of keep that momentum going throughout? Cosmin Nicolaescu (00:40:42) : I think it has to do with finding the right champions at the right levels. So the thing that you normally hear with enterprise sales is find the CEO, they give you buy in, and then you can make decisions. That doesn't really work well, because the managing partner or the CEO at an accounting firm is not going to be close enough to the details to be involved in that. A lot of times they end up being involved towards the end, because they see the promise of Accrual across all these business sides, and so it escalates to their level. But all they can do is say, I want to deploy technology across my firm, go forth and figure out how to do it. And so then it's typically, how do I involve the right level of seniority to have the unblocking escalation path is needed. For us, if I think about our past year, it's generally been the person running tax, like the tax partner at that firm. Because they get to make the decisions for their service line. Then, if they have a technology team, how do you get them involved? And the thing that they care about is, is it secure? Is it easier to integrate with? Is this going to be a pain for me or not? And so we try to be very clear about what we do, and the fact that we've built developer products, I think that helps a lot in how we've approached things. Then finally, it's having people two or three levels down in the organization that are actually doing the work, they are excited about this. So the manager, senior manager kind of level, where they're senior enough to know the organization and experience and have credibility. But not that senior that they're not doing real work. And once you have that combination, then they can get folks around them to be excited and again, keep that number relatively small. Because you don't need to have a hundred people be excited about it. You just need a handful of people and that projects more broadly and that is in the pilot stage. I would say post pilot, in the rollout stage, because I think that is the even harder piece, is having the right folks to drive change, like project managers, and working very closely with them. So you hear about forward deployed engineers in enterprise very commonly now, and I very strongly believe in that. But from the theoretical perspective of having folks on your team very deeply involved in driving the change, not necessarily that I send a person and I ship them and they're going to make all the changes. If I send you an engineer, an accountant, whatever, in your firm, no one knows them. They're not going to understand the dynamics of different teams and people at offices. They're not going to be successful telling people what to do. But basically having someone in that role that works very closely and deeply embedded with the firm on the change management is critical for actually rolling that out and being smooth taxes over tax season. Pablo Srugo (00:43:26) : The other piece that I want to talk to you about is something that you mentioned to me earlier, which was, obviously you have an engineering mindset. Your brain thinks that way, and you've taken that and applied it throughout the whole company that you're building, not just around how you ship software. And my understanding is you're only twenty people today, even though you've raised $75 million. Cosmin Nicolaescu (00:43:50) : 21st person that started last week. Pablo Srugo (00:43:52) : There you go. So tell me more about that. In other words, how are you bringing the engineering mindset to the rest of the organization? Why are you twenty people and not fifty, given what you've raised and given how big the opportunity is? And let's dive into your thinking around talent, recruiting, and operations. Cosmin Nicolaescu (00:43:58) : Going back to your earlier question about how we made choices and what we end up doing, a big factor for us was cultural. What is the culture that we want to create and that we can create? And a few things that we had on our list. One, we wanted to build something where you don't need to have a massive team and this is both pre AI in general. We both enjoy working with smaller groups of people and when you have smaller groups of people, everybody knows everyone. It's just a very different feeling. If I think about early days at both Stripe and Brex, it felt very different and the reason why a lot of the early folks at these rocket ships end up leaving is because the company grows and they lose touch, it becomes much more bureaucratic, and they don't know everyone and just lose that connection. But also, especially in a world in which we believe individual productivity increases so much more, the coordination costs that you had already even before, that's the downside. As a large organization, as a whole, you can do a lot more than a small team, but individually you do less. Each individual is less productive because you have more layers, more coordination necessary, and so you're generally slower. But the sum of it ends up being faster, is the theory. That cost of coordination is higher and higher and higher as the individual productivity increases so much more. There used to be the 10x engineer before. Now I think those 10x engineers are 100x engineers. And so I'd rather have fewer 100x engineers than those 100x engineers not being able to become 100x engineers because of the coordination costs required. And so we are very mindful about every single person that we add in the company, and how, in the long run, we believe we can build Accrual to be a much smaller company than you would typically think. And I've gotten this comment that you just made from several folks. It's like, oh, you guys should be 10, 50, 100 people by now. And I'm like, absolutely not. I very strongly disagree. And I think we would be in a significantly worse place from a lot of different reasons. And that trickles into how you run the company. So we don't have management at this company. I think it's technically everybody reports to me, but we don't do one-on-ones. We don't have scheduled meetings. We trust individuals to be very mature. Is it fully remote? We are remote. So we have folks in SF, New York, Seattle, LA, and then a few other random cities where we hire. But basically, as long as you're in the US and you're willing to travel — because we spend a lot of time in person despite the fact that we're remote — great, that works. We hire generally experienced people. Most folks that we've hired, we've worked with at Stripe, they've had roles, jobs, have seen scale. And the reason for that is not because I have anything against more junior people, it's because I think they require more support around them to grow. It's awful to be a new grad and to work in a remote company. You just don't learn as much. And so we try to optimize for the culture that we want to have. Number three is, there are very loose roles. You're either building the product, growing the product, or enabling people to do one of those two things. That's kind of it. We have very loose lines between people who are selling the product or building stuff nowadays. Everybody uses Claude Code. We've built a lot of internal infrastructure with agents that can autonomously build things and analyze things and tie things together across the company. And everybody in the company uses these tools on a daily basis quite highly. The other week I was looking at some token users, which — I'm not a fan of token maxing by any means — but it was not an engineer. And it's not because they built a bunch of stuff. It's just so much analysis and tying things together. And the independence that people have as a result is insanely high right now. And by doing this, it enables us to have a lot of flexibility. We move people around every few months. If you're an engineer, you're not like, I'm an engineer on the client portal. Maybe you work on a client portal for one or two sprints, or a few weeks, maybe a couple months, and then you're going to move on to something else, either because you want to or because we're going to tell you to do it, such that people get out of their comfort zone and learn the code base across the board. And it is likely slower in the short run, as we're constantly optimizing for the long run. If we assume we're going to be successful in it for the long run, how do we build a company from the ground up with that mindset? And how do we make the technical decisions as well from the ground up with that mindset? Pablo Srugo (00:48:21) : It's interesting, I wonder if it's scalable or not, but another thought that comes to my mind is, by moving people like that you get rid of the issue of protective silos, right? Where it's like, okay, you're on this team doing this thing, and that means this team doing this thing better continue to exist and be important for a long time, and I've got to find a way to make this thing continue to be important, right? Because it's like — Cosmin Nicolaescu (00:48:41) : You're much more mindful, because the code that you're writing, that your agents are writing and you're reviewing realistically — you're going to maintain someone else's code, someone else will maintain your code. There's no deep domain ownership there, so I very much agree with that. And to your point of scalability, we do work trials, for example, where we spend several days with people, typically a couple of days with every hire that we haven't worked with. Part of it is, I don't know how to interview in the day of AI, because if I just give you a problem, we're going to give it to an agent and we'll solve it. And part of it is because I want to see what it's like to work with you. And vice versa, you should see what it's like to work with us. We try to get a bunch of people in person that they would work with and see if you like the culture and if it gets you excited, see the domain and understand it and help you out. We could not do that if we were hiring a lot of people. That is not a scalable process. But if we hire a person a month, much more scalable. You can pay that cost. Pablo Srugo (00:49:41) : And I was going to ask, talking about that, if that's the plan, why raise $75 million? What's the money being used for? Cosmin Nicolaescu (00:49:47) : Part of it is how you actually manage your cap table, and how you think about dilution, and how you think about getting investors to have substantial capital involved. If you get any of the large funds to go and invest a few million dollars, even if you end up being really good, it's just not enough dollars. If I tell you I'm going to take your $5 to a 10,000x exit, you're going to be like, okay, cool. But if you're able to put $50,000 back to a 10,000x exit, very, very different math, to use much smaller numbers. Pablo Srugo (00:50:17) : So part of it is that. In other words, there's value in getting the right investors on board, and those big investors with big funds need to put in big dollars. Cosmin Nicolaescu (00:50:24) : Correct. General Catalyst for us has been invaluable. And in general, both H&R Block and Armanino, they were the ones who introduced us to them. Having a top-tier investor is so valuable. And we've been fortunate to know a lot of the top-tier investors in the industry at Stripe and Brex as well, which I'm very grateful for. The second piece is, it's very hard for me to predict the future. It's very hard for me to predict how the macroeconomy will work and how easy fundraising is or not. There's cycles in and out. It's very hard for me to know what I should spend money on. Everybody last year was talking about the cost of tokens going to zero. And I actually took the opposite side last year, because at the end of the day, tokens represent capacity, and we're very constrained on capacity. We don't have enough energy, we don't have enough data centers, and the demand is exponential. And so if I think about it from that angle, I think actually token costs will probably increase, which is what's happening today. And so I got asked this question at a panel a couple of months ago, which is, if you project out maybe a year or two from now, do you think you're going to spend more on humans or tokens? And I don't think the answer is obviously humans. I think there's actually a good chance you spend more on tokens than you spend on humans. Pablo Srugo (00:51:38) : Are you also finding that the spend on humans, on a per-human basis, is much higher than it used to be? Cosmin Nicolaescu (00:51:43) : We took the approach of — again, not probably scalable — but everybody at Accrual has the same salary. Pablo Srugo (00:51:48) : That's wild. I've never heard of that before. That's crazy. Cosmin Nicolaescu (00:51:50) : So Sid and I make less, because I think it's only fair for founders to be in that situation. But everybody else makes the same amount of money. Because, one, again, the roles are very fluid. I go in and optimize for that. Two, we're hiring pretty senior folks, pretty mature folks. And so I don't want to have different levels and all that. I don't care about it. At the end of the day, you're in this function and you're doing the best you can and helping the company as much as possible. Pablo Srugo (00:52:20) : But does that mean some people come in and they get a huge pay raise relative to what they were making, some people taking a huge pay cut? Or are you just hiring from the top of the mountain anyways? How's that worked out? Cosmin Nicolaescu (00:52:29) : So in general, I would say we've seen both pay raises and pay cuts. I would say most are pay cuts, given where we've hired from, but they're typically insignificant. So everybody makes the same amount of money, but we're pretty generous in that amount to a certain point. And the thinking that we had there was, if we're going to hire senior people, generally they will have higher costs, higher expectations, higher market opportunities. We definitely believe that at this stage of the company, the reason why you're here from a financial perspective should be the equity, not the cash. Otherwise, don't join an early-stage company. It's not worth it. If you don't think that early-stage company will 10x, 50x, 100x, whatever, something very, very substantial, you should not be at an early-stage company. Go work at a big company. You'll have a much better quality of life relative to income. At the same time, there are costs associated with that. And so we don't believe that you should make huge sacrifices. And if you do, you're not going to be able to hire people that have kids, that have a house, a mortgage, certain things. And most people that we have in our company are in that mix. The other thing I tell people is, the value of your equity today is zero. It doesn't matter what our valuation is. Either we go to zero or we grow significantly more. What the value is today doesn't really matter. So I don't do the math in terms of what the dollar value is. I do the math on what's your ownership? What is the ownership in your company? And we give everybody quite substantial ownership in the company. Because, again, if we're not hiring a lot of people, we can afford to be much more generous with the people that we do hire from a cap table perspective, without being overly dilutive to everybody else. So all these things trickle in and can compound one way or the other. And basically, with every person that we hire, I walk through it with them. Here's the ownership that you'll have in this company. What do you believe this company will be at? Here's where I believe it will be at within a reasonable time frame, like four to five years, not 50 years. Within this stock grant, where do we believe each of us — what are we modeling on our side? Here's with your numbers what it looks like. Here's with my numbers what it looks like. Here's the average, if there's a delta, et cetera. And getting a sense of where this could be. And that's what should excite people, and that's what people should work for, versus high cash guarantee, no risk. And we've had people that said, I need more cash. And respectfully, I tell them that then you shouldn't be at an early-stage company. You don't need to be at a 20-person startup. There's nothing wrong with saying, I can't take this risk, or I can't afford to take this risk. Then go look at a bigger company that will have very different equity packages and different cash packages. I think it's a very good filtering function, which I think we find around equilibrium — between not super low, where we can hire people. We've been very successful at converting folks that do successfully in our work trial and filtering out people that probably are not fit for the super early stage. Pablo Srugo (00:55:17) : Yeah, it gives you an edge, this idea that you're hiring fewer, because especially on the stock grants, it's hard for somebody that's going through the more normal hiring path — let's say another scaling startup that raised $50 million — to be able to match those equity grants, just by the sheer number that they're going to have 50 people, 100 people instead of 21. Cosmin Nicolaescu (00:55:34) : Because either you're going to say all these companies give the same grants. And if you do that, you're just going to dilute a lot more. So it's one thing to say I'm giving someone 1% of the company, but it's a different thing of what is that 1% in four to five years? That's right. Dilution is something that can hit people quite heavily. And so that's why I do think you can be much more generous if you have fewer people. Pablo Srugo (00:55:58) : Perfect. So listen, let's stop it there. I think it's been a great episode. Maybe just as a last question, what would be the number one piece of advice that you'd give, or intend to give, maybe early stage founders that are in this Product Market Fit phase? Cosmin Nicolaescu (00:56:10) : I would say the most successful founders that I've seen, and the thing that I try to emulate from Patrick and John and other founders that I've met throughout my life, is long term conviction. Things compound over time. It's very easy to be dealing with the problems of today because as an early stage founder, pre-product market fit, right after product market fit, you don't have that escape velocity, which we don't. You are obsessed about, will I exist in the next week, the next month? That's the kind of time frame that you think through. And I think that is right to focus on that. But the things that move the needle in the long run are the things that compound over time. And so having very high conviction over a small number of things and letting those things compound is what I think makes very big businesses succeed over time and thrive, as long as you get those things right. I do think one of my favorite operating principles is from Amazon, which is that good leaders often are right. I think there's something very, very meaningful there. You have to be right, and if you're not, then you probably shouldn't be a founder. Pablo Srugo (00:57:13) : Well, Cos, thanks so much for spending the time with us, man. It's been great. Cosmin Nicolaescu (00:57:13) : Thanks for having me, man. Pablo Srugo (00:57:13) : 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.