They Said No—Now They're Our Biggest Customer

"We don't need your services." Those were the words that greeted Jordan Katz and Michael Mattheakis when they first pitched their insurance automation idea to one of the industry's largest players.

Fast forward three years, and that same company is now one of Comulate's largest customers. The AI-powered insurance platform has scaled to over 8 figures in ARR, proving that sometimes the best opportunities lie exactly where everyone tells you not to look.

In this candid interview, the Comulate co-founders reveal how they discovered a $10+ million opportunity in an industry they knew nothing about, why being outsiders became their greatest advantage, and how they're fundamentally reimagining entire business functions rather than just optimizing existing workflows.

Jordan: By the time I was in college, I was starting to try to build software in a software company and spent thousands of hours building actually one product in particular. In reality, I realized as I was going, I was building a huge dumpster fire. The platform had 500 features. Each of them was like 5% built.

Through that, I realized how actually complex it was to build great software, and I realized I needed to learn how to do that. I kind of came across and learned about the role of product management, and I read this book called "Cracking the PM Interview." The author was Jackie, who ran product at Asana. It was just like 20 pages into it, I was like, I need to work for Jackie, and that's how I ended up going to Asana with the mission of I want to learn how to build great product from Jackie.

I met my co-founder Michael towards the end of college. I was going to work in product at Asana, and Michael reached out because he was interested in understanding what the product was like. We met for lunch and it was one of those really 1 in a million conversations where you meet someone and think like this person is different. It sounds pretty crazy, but at that moment, I was actually already thinking like this might be someone I start a company with. Unfortunately, I think for me, Michael didn't feel that way. He was going to this company that no one knew about called Brex, and he was going to be the 10th hire there.

Michael: I was an engineer at Brex prior. I joined the company out of college when it was just about 10 people and saw it grow super quickly to over 900 during my time there.

Something the Brex founders are really incredible at that I picked up on is being able to see the business at both a 10,000 ft view and a 10-foot view. Even from the very beginning, they were thinking about, hey, how do we scale this to being a 100+ billion dollar company at some point and becoming the global substrate for financial money movement. At the same time, they're extremely good at saying, hey, what growth hacks do we need to do this week to be able to push the company forward, what next feature do we need to build and what should be the design and product mocks to be able to do that.

Those two skill sets, it's pretty rare to find in the same person, and that's a big part of what I found to be really successful in that early stage is being able to do both. It was a pretty incredible first job out of college, but also I had always thought about starting a company. Jordan and I were super good friends, and we had been talking about it more and more over time. Jordan had just left his job, and it got to a point where we were starting nights and weekends to work on different startup ideas together. All of these factors seemed to come together. It said like, hey, there was a moment to do this, now is definitely it.

Jordan: When Michael and I decided we're gonna do this together, we actually didn't know anything about insurance or accounting. From our years at Asana and Brex, we had a few ideas that we were excited about, but as we went through the early process of validating those, talking to the folks that we would try to serve and would be customers, the problems that we thought were worth solving weren't actually in practice, they didn't feel as big as it seemed, and the existing solutions didn't seem as bad as we thought they were.

Michael and I really got into the idea of wanting to build something that didn't already exist. We had started learning about insurance and we were immediately stunned by the amount of paper, spreadsheets, and time spent copying information from the analog world to the digital world of very legacy software and back. We were really especially fascinated by the accounting workflows. We saw that there were many intermediaries in the value chain that sort of bubbled up into this observation: this whole industry has an additional extra version or extra type of accounting that other industries don't have to deal with, and there's no good solutions that automate that. Maybe we can develop that solution that ultimately led to the foundation for Comulate.

Jordan: People we reached out to early on in the industry were quite skeptical. It was hard initially to talk to people in this industry, and you relied on the few 2nd or 3rd degree connections you would have to get a conversation, and then the first thing I hear is, "You know, I just want to tell you, I'm taking this call as a favor, but we don't need your services. We have solved for this already. We're all set. I'm just trying to do this as a favor."

And at the time, that was a massive blow to us. We're like, oh wow, this is one of the largest companies in the space. If they don't need this, why would anyone else need it? But then as we actually started diving into it, talking to more people, we kind of got the sense like, hey, he might just not be exposed to this specific problem within his company. We got more and more evidence, many others actually did need our services.

What we spent a ton of time on in the early days was just on Zoom calls with our customers, really, really deeply understanding what they're doing. Part of what helped us is we weren't salespeople. We were product people trying to solve a problem and build a solution for these folks. So our questions usually were very deep about their workflows, about edge cases, about why something was done a certain way in the first place.

Jordan: I remember one of our early customers wrote me a note when they retired, and they had been instrumental in taking the bet on us in one of those special relationships. They had said "50 years of working at this company, working with you is the first time someone who's brought in a product has ever felt like they understand us." To me that kind of indicates the strength of coming in as a product person versus a sales-oriented mind of being able to build trust and build the depth of knowledge and domain expertise that lends itself to a customer wanting to believe you can actually solve their problem versus make promises.

Luckily for us, there were just enough special people at a few companies who sort of saw some potential and spent the time and were willing to help us build the initial kernel. It's somewhat luck and finding the right person at the right time, it resonating on that day and then that conversation going well. Luckily, you don't need that many of those experiences and relationships at this point if you can really develop some real core relationships.

That company that told us "we don't need your services" is now one of our largest customers. Even though at the moment it feels like this setback and you're not sure of the viability of what you're doing and whether it's even worth pursuing anymore, if you just keep on persisting and actually getting more information, building up that bottoms up confidence of what you're working on, it actually led us to persist past that.

Jordan: Many people are thinking about AI in terms of how can you use it to improve existing workflows, and this is especially true in traditional industries where it's like, how can you now automate something you couldn't previously automate with the prior generation of tools, or how to augment humans by writing an email or navigating a website. But I feel like that's not actually the most interesting question.

I think the real question is about what fundamental constraints created these workflows in the first place and do those constraints actually still exist? If you think about any prior platform shift like web to mobile in the 2010s, the most successful products were actually rethinking what should be on a mobile app, for example, like Instagram, as opposed to a lot of the unsuccessful ones that just took a normal website and put it into a mobile app, and it actually wasn't that compelling. It's because there are actually new things that are possible, right? You can drag, you can multitask, etc.

Now, with the AI platform shift, it's a similar situation. In this context, it's all based on this foundation of really understanding what are the fundamental problems that customers are facing. When customers first come to us, we'll map out their entire process and it often looks like this crazy flow chart with 30 boxes all interconnected. A big part of what we do is look at that and say, hey, you can actually simplify this to 3 or 4 steps, especially with what software and AI can do, and this is how our product can actually do that.

Jordan: That step of actually simplification rather than just trying to fit the crazy complexity into a product is what actually lets us rethink and reimagine what's possible. That's the difference between optimizing what exists today versus reimagining an entire function from scratch. The latter is a completely different opportunity that I think the winners of that transition will have in common.

Jordan: A few years ago when we started the company, building software for vertical industries, especially automating work in them, was not that popular. It's a pretty unconventional thing to go into something like the insurance industry as outsiders and try to understand and automate a lot of the work that's happening there, but now it's super hot. So 70% of the most recent YC batch is now doing AI agents for X industry. I think a big part of that is just because where AI is having the most impact right now are these places that have been pretty untouched by software historically.

I think the biggest advice I have is look where others don't want to look. There's so many reasons that will give you to not explore a space and frankly, if we had tried to raise our seed round not doing any discovery or validation before, we would have never done this because everyone told us it was a bad idea. But going in the field and sort of doing what folks have told you is not a good place to play in, sometimes that's where the opportunity is. I really think it's important to go on a personal mission to find the truth because you never really know their altitude and understanding of a problem until you dive in and try to figure it out yourself.

Jordan: What has me really excited about the future is there's still hundreds of billions of dollars spent on manual work every year in this space, and AI has opened that up into a massive opportunity, and we're right at the forefront of being able to take advantage of that. That means that to me this is very obviously one of the largest opportunities, and you think about the next 10 and 20 years in software. That's super exciting for our team.