Transcriptly

How to Find $10M+ Ideas in Markets We’d Never Worked In | Comulate, Jordan Katz & Michael Mattheakis

EO12:49Added Aug 31, 2026

They built a multi-million dollar AI company by going against Silicon Valley's playbook. Jordan (CEO) and Michael (CTO) are the co-founders of Comulate — a startup that’s quietly transforming the insurance industry using AI.

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. Intro

  2. 00:00So there's a popular belief in Silicon
  3. 00:02Valley of the best way to find startup
  4. 00:04ideas is to try to solve your own
  5. 00:06problems. But if everyone in the tech
  6. 00:09industry is following that path, it
  7. 00:11leads to this saturation of problems for
  8. 00:14folks in the tech industry. That's what
  9. 00:16led us to the insurance industry quickly
  10. 00:17realized there's unsolved problems
  11. 00:19everywhere. People we reached out to
  12. 00:22early on in the industry were quite
  13. 00:23skeptical. The f first thing I hear is,
  14. 00:26you know, I just want to tell you, you
  15. 00:28know, I'm taking this call as a favor,
  16. 00:31but we don't need your services. We have
  17. 00:33solved for this already. We're all set.
  18. 00:35I'm just trying to do this as a favor.
  19. 00:37And in retrospect, it's like, why would
  20. 00:39they talk to two random outsiders?
  21. 00:41People are busy and hear pitches like
  22. 00:43that all the time. We had that quote in
  23. 00:45our old office on the wall of that
  24. 00:47formative moment of we don't need your
  25. 00:49services. And that company that told us
  26. 00:51we don't need your services is now one
  27. 00:54of our largest customers.
  28. 00:58I'm Jordan. I'm the co-founder and CEO
  29. 01:00of Commulate. We're transforming the
  30. 01:02insurance industry with AI. 3 years ago,
  31. 01:04it seemed crazy to most folks that we
  32. 01:06wanted to focus on the back office
  33. 01:09insurance workflows. When you have like
  34. 01:10two people with no background in doing
  35. 01:12the thing that they're saying they're
  36. 01:14going to be able to automate for you,
  37. 01:16who the heck's going to take those those
  38. 01:18people seriously? But since then, we've
  39. 01:20scaled to over eight figures of ARR and
  40. 01:22we're just getting started.

  41. How to Know Someone’s the Right Co-founder

  42. 01:32By the time I I was in college, I was
  43. 01:35starting to try to build software in a
  44. 01:37software company and spend thousands of
  45. 01:39hours building actually one product in
  46. 01:40particular. In reality, I realized as I
  47. 01:43was going, I was building a huge
  48. 01:44dumpster fire. Platform had 500
  49. 01:46features. is each of them was like 5%
  50. 01:48built. Through that I realized like how
  51. 01:50actually complex it was to build great
  52. 01:52software and I realized I think like I
  53. 01:55needed to learn how to do that and I
  54. 01:56wanted to learn how to build great
  55. 01:58software and I kind of came across and
  56. 01:59learned about the role of product
  57. 02:00management and I read this book
  58. 02:02particularly called cracking the PM
  59. 02:04interview that sort of several folks had
  60. 02:05told me to read the author was Jackie
  61. 02:07who ran product at a sauna it was like
  62. 02:1020 pages into it I was like I need to
  63. 02:13work for for Jackie and that's how I
  64. 02:15ended up going to a sauna with a mission
  65. 02:17of I want to learn how build great
  66. 02:19product from Jackie and then I ended up
  67. 02:20being able to learn from a lot of other
  68. 02:22great folk around that. I met my
  69. 02:24co-founder Michael towards the end of
  70. 02:25college. I was going to work in product
  71. 02:27at ASA and Michael reach out cuz he was
  72. 02:29interested in in understanding what
  73. 02:31product was like. We met for lunch and
  74. 02:32it was one of those really one ina-
  75. 02:36million conversations where you meet
  76. 02:38someone and think like this person is
  77. 02:39different. And it sounds pretty crazy,
  78. 02:41but at that moment, I was actually
  79. 02:44already thinking like, this might be
  80. 02:47someone I start a company with.
  81. 02:49Unfortunately, I think for me, Michael
  82. 02:51didn't feel that way. He was going to
  83. 02:53this company that no one knew about
  84. 02:55called Brex, and he was going to be the
  85. 02:5610th hire there. I'm Michael. I'm the
  86. 02:58CTO and co-founder of Commulate. I was
  87. 03:00an engineer at Brax uh prior. I joined
  88. 03:03the company out of college when it was
  89. 03:05just about 10 people and saw it grow
  90. 03:07super quickly to over 900 during my time
  91. 03:09there. Something the Bra founders are
  92. 03:12really incredible at that I picked up on
  93. 03:15is being able to see the business at
  94. 03:17both a 10,000 ft view and a 10-ft view.
  95. 03:20Even from the very beginning they were
  96. 03:22thinking about, hey, how do we scale
  97. 03:24this to being 100 plus billion dollar
  98. 03:26company at some point and becoming like
  99. 03:28the global substrate for financial money
  100. 03:31movement. At the same time, they're
  101. 03:33extremely good at saying, hey, what
  102. 03:36growth hacks do we need to do this week
  103. 03:38to be able to push the company forward?
  104. 03:40You know, what next feature do we need
  105. 03:42to build and what should be the design
  106. 03:43and product mocks to be able to do that?
  107. 03:46And those two skill sets, it's pretty
  108. 03:48rare to find in like the same person.
  109. 03:50And that's a big part of what I found to
  110. 03:52be really successful in that early state
  111. 03:54is being able to do both. It was a
  112. 03:56pretty incredible first job out of
  113. 03:58college. But also, I had always thought
  114. 04:00about starting a company. Jordan and I,
  115. 04:03like I mentioned, we were super good
  116. 04:04friends and we had been talking about it
  117. 04:06more and more over time. Jordan had just
  118. 04:08left his job and it got to a point where
  119. 04:11we were starting like nights and
  120. 04:12weekends to work on different startup
  121. 04:14ideas together. All of these factors
  122. 04:16seem to come together. It said like,
  123. 04:18"Hey, if there was a moment to do this,
  124. 04:20now is definitely it." So, I leave Bra
  125. 04:23and jump right into starting a company
  126. 04:24with Jordan.
  127. 04:26[Music]

  128. How to Break Into an Unfamiliar Industry

  129. 04:34When Michael and I decided we're going
  130. 04:36to do this together, we actually like
  131. 04:38didn't know anything about insurance or
  132. 04:40accounting. From our years at Asana and
  133. 04:42Brex, we had a few ideas that we were
  134. 04:45excited about. But as we went through
  135. 04:46the early process of validating those,
  136. 04:48talking to the folks that we would try
  137. 04:50to serve and would be customers, the
  138. 04:52problems that we thought were worth
  139. 04:54solving weren't actually in practice,
  140. 04:57they didn't feel as big as it seemed and
  141. 04:59and the existing solutions didn't seem
  142. 05:01as bad as we thought they were. And I
  143. 05:02think Michael and I really got into the
  144. 05:05idea of wanting to build something that
  145. 05:08didn't already exist. We had started
  146. 05:09learning about insurance and we were
  147. 05:11immediately stunned by the amount of
  148. 05:13paper, spreadsheets and time spent like
  149. 05:16copying information from the analog
  150. 05:18world to the digital world of very
  151. 05:19legacy software and back in that we were
  152. 05:22really especially fascinated by the
  153. 05:24accounting workflows. We saw that there
  154. 05:26was you know many intermediaries and in
  155. 05:28the value chain that sort of bubbled up
  156. 05:30into this observation. This whole
  157. 05:32industry has an additional extra version
  158. 05:34or extra type of accounting that other
  159. 05:36industries don't have to deal with and
  160. 05:38there's no good solutions that automate
  161. 05:40that. Maybe we can develop that solution
  162. 05:42and that ultimately led to the kind of
  163. 05:44foundation for commulate.
  164. 05:47People we reached out to early on in the
  165. 05:49industry were quite skeptical. It was
  166. 05:52hard initially to like talk to people in
  167. 05:54this industry and you relied on like the
  168. 05:56few second or third degree connection
  169. 05:57you would have to get a conversation and
  170. 06:00then the first thing I hear is you know
  171. 06:02I just want to tell you you know I'm
  172. 06:04taking this call as a favor but we don't
  173. 06:07need your services we have solved for
  174. 06:09this already. We're all set. I'm just
  175. 06:11trying to do this as a favor. And at the
  176. 06:14time that was a massive blow to us.
  177. 06:15We're like oh wow like this is one of
  178. 06:17like the largest companies into space.
  179. 06:19if they don't need this, like why would
  180. 06:21anyone else need it? But then as Baka
  181. 06:23started diving into it, talking to more
  182. 06:26people, we kind of got the sense like,
  183. 06:27hey, he might just not be exposed to
  184. 06:29this specific problem within his
  185. 06:31company. And we got more and more
  186. 06:33evidence many others actually did need
  187. 06:35our services. And what we spent a ton of
  188. 06:38time on in the early days was just on
  189. 06:41Zoom calls with our customers really,
  190. 06:43really deeply understanding what they're
  191. 06:45doing.
  192. 06:45And part of what helped us is we weren't
  193. 06:48salespeople. We were product people like
  194. 06:50trying to solve a problem and build a
  195. 06:52solution for these folks. So our
  196. 06:53questions usually were like very deep
  197. 06:55about their workflows, about edge cases,
  198. 06:57about why something was done a certain
  199. 06:59way in the first place. I remember one
  200. 07:02of our early customers wrote me a note
  201. 07:04when they retired because I was issuing
  202. 07:06them at retirement and they had been
  203. 07:07instrumental in taking the bet on us in
  204. 07:09one of those special relationships and
  205. 07:10they had said 50 years of working at
  206. 07:13this company working with you was the
  207. 07:14first time someone who's brought in a
  208. 07:16product has ever felt like they
  209. 07:18understood us and to me that kind of
  210. 07:21indicates the strength of coming in as
  211. 07:23product person versus a salesoriented
  212. 07:26mind of being able to build trust and
  213. 07:29and build the depth of knowledge and
  214. 07:31domain expertise that lends itself to a
  215. 07:34customer wanting to believe you can
  216. 07:36actually solve their problem versus make
  217. 07:38promises. Luckily for us, there was like
  218. 07:40just enough special people at a few
  219. 07:43companies who sort of saw some potential
  220. 07:45and spend the time and were willing to
  221. 07:47sort of help us build the initial
  222. 07:49kernel. It's somewhat luck and finding
  223. 07:51the right person at the right time, it
  224. 07:53resonating on on that day and then that
  225. 07:55conversation going well and a lot
  226. 07:57happens through that. Luckily, you don't
  227. 07:58need that many of those experiences and
  228. 08:01relationships at this point if you can
  229. 08:03really develop some real core
  230. 08:05relationships.
  231. 08:07And that company that told us we don't
  232. 08:09need your services is now one of our
  233. 08:11largest customers. Even though at the
  234. 08:13moment it feels like this setback and
  235. 08:15you're not sure of the viability of what
  236. 08:18you're doing and whether it's even worth
  237. 08:20pursuing anymore, if you just keep on
  238. 08:22persisting and actually getting more
  239. 08:24information, building up that bottoms up
  240. 08:28like confidence of what you're working
  241. 08:29on, it actually led us to persist past
  242. 08:32that they ended up needing the services.

  243. Applying AI to Solve the Right Problem

  244. 08:43Many people are thinking about AI in
  245. 08:46terms of like how can you use it to
  246. 08:48improve existing workflows and this is
  247. 08:51especially true in traditional
  248. 08:53industries where like how can you now
  249. 08:55automate something you couldn't
  250. 08:56previously automate with the prior you
  251. 08:58know generation of tools or how to
  252. 09:00augment humans by writing an email or
  253. 09:02navigating a website. But I feel like
  254. 09:04that's not actually the most interesting
  255. 09:05question. I think the real question is
  256. 09:07about like what fundamental constraints
  257. 09:10created these workflows in the first
  258. 09:12place and do those constraints actually
  259. 09:14like still exist? If you think about any
  260. 09:17prior platform shift like web to mobile
  261. 09:20in the 2010s, the most successful
  262. 09:23products were actually rethinking what
  263. 09:26should be on like a mobile app for
  264. 09:28example, like Instagram for example, as
  265. 09:30opposed to a lot of the unsuccessful
  266. 09:32ones just took a normal website and put
  267. 09:34it into a mobile app and it actually
  268. 09:36wasn't that compelling. It's because
  269. 09:37there's actually like new things that
  270. 09:39are possible, right? You can like drag,
  271. 09:40you can like multitask, etc. Now with
  272. 09:43the AI platform shift, it's a similar
  273. 09:46situation. In this context, it's all
  274. 09:48based on this foundation of like really
  275. 09:50understanding what are like the
  276. 09:52fundamental problems that customers are
  277. 09:54facing. When customers first come to us,
  278. 09:56we'll map out their entire process. And
  279. 09:58it often looks like this like crazy
  280. 10:01flowchart with like 30 boxes all
  281. 10:03interconnected. A big part of what we do
  282. 10:05is look at that and say, "Hey, you can
  283. 10:07actually simplify this to three or four
  284. 10:10steps, especially with what software and
  285. 10:12AI can do. And this is how our product
  286. 10:14can actually do that." And that uh step
  287. 10:17of actually simplification rather than
  288. 10:20just trying to fit the crazy complexity
  289. 10:24into a product is what actually lets us
  290. 10:26able to rethink and reimagine what's
  291. 10:29what's possible. That's the difference
  292. 10:31between optimizing what exists today
  293. 10:34versus like reimagining an entire
  294. 10:36function from scratch. The latter is a
  295. 10:38completely different opportunity that I
  296. 10:39think will the winners of that
  297. 10:41transition will have in common.
  298. 10:43A few years ago when we started the
  299. 10:44company building software for vertical
  300. 10:46industries like automating work in them
  301. 10:49was not that popular. It was a pretty
  302. 10:52unconventional thing to go into
  303. 10:53something like the insurance industry's
  304. 10:54outsiders and try to understand and
  305. 10:57automate a lot of the work that's
  306. 10:58happening there. But now it's super hot.
  307. 11:00So like 70% of the most recent YC batch
  308. 11:03is now doing AI agents for X industry. I
  309. 11:07think a big part of that is just because
  310. 11:10where AI is having the most impact right
  311. 11:12now are these places that have been
  312. 11:14pretty untouched by software
  313. 11:17historically. I think the biggest advice
  314. 11:19I have is like look where others don't
  315. 11:22want to look. There's so many reasons
  316. 11:23that folks will give you to like not
  317. 11:26explore a space. And frankly, if we had
  318. 11:28tried to raise our seed round, not doing
  319. 11:30any discovery or validation before, we
  320. 11:32would have never done this because
  321. 11:33everyone told us it was a bad idea. But
  322. 11:35going out on the field and sort of doing
  323. 11:36what folks have told you is not a good
  324. 11:39place to play in. Sometimes that's where
  325. 11:41the opportunity is. And I really think
  326. 11:43it's important to like go on a personal
  327. 11:45mission to find the truth because you
  328. 11:46never really know their altitude and
  329. 11:48understanding of a problem until you
  330. 11:49dive in and try to figure it out
  331. 11:51yourself. So what has me really excited
  332. 11:52about the future is there's still
  333. 11:55hundreds of billions of dollars spent on
  334. 11:58manual work every year in this space and
  335. 12:00AI has opened that up into a massive
  336. 12:03opportunity and we're right at the
  337. 12:05forefront of being able to take
  338. 12:06advantage of that
  339. 12:08and that means that to me this is very
  340. 12:10obviously one of the largest
  341. 12:12opportunities and you think about the
  342. 12:14next 10 and 20 years in software that's
  343. 12:16super exciting for our
  344. 12:18[Music]
How to Find $10M+ Ideas in Markets We’d Never Worked In | Comulate, Jordan Katz & Michael Mattheakis — Transcriptly