Transcriptly

The Fastest Way to Know if Your Product Market Fit Is Real | Serval CEO, Jake Stauch

EO13:40Added Aug 31, 2026

Jake Stauch is the co-founder and CEO of Serval, an AI-native platform that automates the work of corporate IT teams, recently valued at $1 billion after a $...

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. Intro

  2. 00:00We had rabid fans of our product that
  3. 00:01loved what we were doing. And that gave
  4. 00:03me this delusion that we had found
  5. 00:05product market fit or we were just
  6. 00:06around the corner from really unlocking
  7. 00:08massive growth. It shouldn't be one good
  8. 00:10customer conversation, three good
  9. 00:12customer conversations. You have to have
  10. 00:14enough where you really feel like you
  11. 00:15understand the problem deeply. You
  12. 00:17should be really ruthless around your
  13. 00:21judgment on whether you found product
  14. 00:23market fit. And it has to be in a
  15. 00:26genuine interest in them and in making
  16. 00:29their lives better that drives that
  17. 00:31relationship forward and gives you the
  18. 00:32insights you need to build a better
  19. 00:34product. I'm Jake. I'm the co-founder
  20. 00:36and CEO of Serville. Servil is an AI
  21. 00:38native platform for IT teams. We help
  22. 00:41automate help desk requests, onboarding,
  23. 00:43offboarding, just in time access
  24. 00:45requests. Basically, the long tale of
  25. 00:47things that it has to do. And we just
  26. 00:49recently announced our $75 million
  27. 00:52series B led by Sequoia at a billion
  28. 00:54dollar valuation. From founding to a
  29. 00:57term sheet valuing us at a billion
  30. 00:59dollars was 18 months.

  31. Why Fake Product-Market Fit Feels Exactly Like Real

  32. 01:11I don't know that I would describe
  33. 01:13myself as a genius, but yeah, I did get
  34. 01:15a a good SAT score in high school. I
  35. 01:17think the way I would think it it
  36. 01:18affected me is it just gave me a lot of
  37. 01:20confidence. I think that that confidence
  38. 01:22is really helpful just knowing that I
  39. 01:24can figure things out. I can learn
  40. 01:26things. I can I can be competitive. Uh I
  41. 01:28I don't have anything to prove. I'm just
  42. 01:30trying to to learn. I was working in a
  43. 01:32neuroscience lab at Duke. But as a kid,
  44. 01:35I think I was just really curious. I
  45. 01:37really like to learn. I really got
  46. 01:39interested in topics very quickly. dove
  47. 01:41in, learned a lot, and then got bored
  48. 01:43and moved on to the next topic and not
  49. 01:45really sticking with anything all the
  50. 01:47way through because I was just kind of
  51. 01:48hopping between interests. And so I
  52. 01:51always wanted to start a company. My mom
  53. 01:53was an entrepreneur and I always
  54. 01:55envisioned myself I tried to start a
  55. 01:57business when I was a teenager and I
  56. 01:59think that was ultimately the goal. And
  57. 02:02even when I was in college, I joined a
  58. 02:03bunch of entrepreneurship groups. I I
  59. 02:05ended up moving into an entrepreneurship
  60. 02:06group. So even though I was studying
  61. 02:07neuroscience, I I was very interested in
  62. 02:10in startups and starting companies. I
  63. 02:12dropped out of college to start my first
  64. 02:14company at a company testing advertising
  65. 02:16with brain scans. At the time, just like
  66. 02:18I'll take a few months off and then I'll
  67. 02:20just come back to school if this doesn't
  68. 02:21take off. And then that first semester
  69. 02:23off became a second semester off became
  70. 02:25a third semester off and then eventually
  71. 02:27became clear that I was never going
  72. 02:28back. And things went well enough. And
  73. 02:31something happened during that process
  74. 02:33where one of our customers said, "Hey,
  75. 02:35can I take that headset home? Can I put
  76. 02:37that on my son and see when he's paying
  77. 02:40attention to his homework and when he's
  78. 02:41not? He has ADHD and I just want to help
  79. 02:44him give get feedback." And that story
  80. 02:46really stuck with me and we decided to
  81. 02:49build kind of a pet project of hey,
  82. 02:50maybe we could build a video game for
  83. 02:52kids with ADHD. And we ended up pivoting
  84. 02:54the company to build out a consumer
  85. 02:56product for kids with ADHD, building new
  86. 02:57hardware, a new headset, and that became
  87. 02:59Neuroplus. ran that company for about
  88. 03:02seven years. We had fans, we had rabid
  89. 03:05fans of our product that loved what we
  90. 03:07were doing. And that gave me this
  91. 03:09delusion that we had found product
  92. 03:10market fit or we were just around the
  93. 03:12corner from really unlocking massive
  94. 03:14growth. But so at the time when I got
  95. 03:15Forbes 730, I didn't think of the
  96. 03:18company as falling apart. In retrospect,
  97. 03:19obviously we were not anywhere near
  98. 03:21product market fit. So maybe we did have
  99. 03:23some product market fit, but it was with
  100. 03:24a market that was so small that it it
  101. 03:27was pretty irrelevant and
  102. 03:28inconsequential. We're just a little bit
  103. 03:30away from unlocking some massive growth.
  104. 03:32But during that time, it's it's ticking
  105. 03:34up. It feels good. It feels like it's
  106. 03:35working. We just done done a successful
  107. 03:38Kickstarter campaign. A lot of people
  108. 03:39are buying our product. There's always
  109. 03:40one more thing that I felt like was
  110. 03:42going to unlock the next stage of
  111. 03:43growth. And we could probably keep it
  112. 03:45going in some form for a very long time,
  113. 03:47but it was not going to be the company
  114. 03:48that I imagined and I envisioned. So,
  115. 03:50probably took a year or two after
  116. 03:52winding down Neuroplus for me to really
  117. 03:54be able to look back and realize that we
  118. 03:56were never that close. It shouldn't be
  119. 03:58one good customer conversation, three
  120. 04:00good customer conversations. You have to
  121. 04:02have enough where you really feel like
  122. 04:04you understand the problem deeply and
  123. 04:06you understand how your solution is
  124. 04:07really going to solve it for a big
  125. 04:09enough part of the market. I think for
  126. 04:11first-time founders, you should be
  127. 04:13really ruthless around your judgment on
  128. 04:17whether you found product market fit.
  129. 04:18And so, obviously, you're always
  130. 04:20starting with early adopters and you're
  131. 04:21always having to like grow into the
  132. 04:22larger mass market. But I think you have
  133. 04:24to be honest with yourself about, you
  134. 04:26know, are you starting with early
  135. 04:27adopters and you actually are going to
  136. 04:29be able to get to the rest of the market
  137. 04:30or is your entire market just a kind of
  138. 04:33a crazy early adopter set and and that
  139. 04:35could be a hard thing to figure out. I
  140. 04:37think some of the signals that we had
  141. 04:39early on was that there there were
  142. 04:41things that were very unique and
  143. 04:42different about our customers that
  144. 04:44didn't look like the rest of the market.
  145. 04:46And I think that that is one signal that
  146. 04:48maybe you're not playing in as big of a
  147. 04:50market as you think is if your your
  148. 04:53early customers all have something in
  149. 04:54common that seems very very strange and
  150. 04:57you don't see a a natural gradient
  151. 04:59towards the rest of the market from
  152. 05:01jumping from those customers to the
  153. 05:02rest. It feels like it's working
  154. 05:04especially if you have a lot of optimism
  155. 05:06which all founders do especially early
  156. 05:08founders early career founders. It's
  157. 05:09hard to look at great customer feedback
  158. 05:12and have the skepticism and the cynicism
  159. 05:16to say I don't think this is working. We
  160. 05:19should do something else. And you
  161. 05:20understand how your solution is really
  162. 05:22going to solve it for a big enough part
  163. 05:24of the market. And then that experience
  164. 05:25of building hardware and software
  165. 05:27platforms is what led me to Vicata
  166. 05:30before leaving to start serving.

  167. The Signal That Tells You It's Working

  168. 05:36I think the thing that sticks with me
  169. 05:37the most is this idea of you have really
  170. 05:40strong market pull and product market
  171. 05:42fit. The rest kind of falls into place.
  172. 05:44I was lucky to have a friend an early
  173. 05:46adviser to Neuroplus that I had gone to
  174. 05:49college with. He had joined Verata as an
  175. 05:51early employee running marketing and he
  176. 05:53initially called my wife and said uh hey
  177. 05:55we need people uh this is crazy. And I
  178. 05:58remember his words were you know how
  179. 05:59everything at a startup is really hard
  180. 06:00and nothing works. and eventually flew
  181. 06:03out on site, walked into the office,
  182. 06:05felt the energy and realized this is
  183. 06:07where I have to be. I have to see what
  184. 06:09this feels like, what this looks like if
  185. 06:10I'm going to be a successful
  186. 06:11entrepreneur in the future. It's only in
  187. 06:13retrospect when I joined Vicata and I
  188. 06:14saw real product market fit that you see
  189. 06:16how different it is. So, when I was on
  190. 06:18first call with a customer, a sales rep
  191. 06:21was demoing the product and I didn't
  192. 06:23think the sales rep did a very good job
  193. 06:25and they messed up a lot. They kind of
  194. 06:28fumbled their way through the demo. They
  195. 06:30got a couple things wrong. And the
  196. 06:31customer was a little bit confused and I
  197. 06:33just thought, "Oh man, this is not going
  198. 06:35well." And then at the end of the call,
  199. 06:37the customer says, "Okay, sounds good.
  200. 06:39Could you send me a quote? I think we'll
  201. 06:41probably buy 30 cameras or so in the
  202. 06:43next month." The biggest question mark
  203. 06:44for us was not whether or not this was
  204. 06:46the right approach. I think it's
  205. 06:47intuitive and obvious that if you could
  206. 06:50do it, if the technology existed, that
  207. 06:51would just be the right approach. And
  208. 06:53you don't have to be perfect at every
  209. 06:55other aspect of the business if you get
  210. 06:56that part right. And so it's always been
  211. 06:58important for us from the very beginning
  212. 06:59is focus on the product experience.
  213. 07:01Focus on making sure you're building
  214. 07:02something customers want and really
  215. 07:04customers love and then the other things
  216. 07:06kind of fall in line. So if you don't
  217. 07:08have product market fit, you can't fix
  218. 07:09the rest of the organization. It means
  219. 07:11that there's an answer to the problem in
  220. 07:13the back of the book is kind of how I
  221. 07:14liken it is so hard and you have to work
  222. 07:16through problems but you know that
  223. 07:18there's an answer to all the problems
  224. 07:19that you're trying to solve. That's what
  225. 07:20I deal with. And if you could solve
  226. 07:22that, that would be meaningful. And that
  227. 07:24was really how Serbal was born.

  228. The Failure That Built a $1B Company

  229. 07:30It was a year into servo before I really
  230. 07:33believed that we were anywhere close to
  231. 07:35product market fit because I was so
  232. 07:36skeptical of every conversation I had.
  233. 07:39We are an AI platform for employee
  234. 07:41support. So when employees go and ask
  235. 07:42questions like they need access to an
  236. 07:44application, they need a password reset,
  237. 07:46they need to request time off or just
  238. 07:47get the Wi-Fi information, Serville
  239. 07:49automates the answers to those
  240. 07:50questions. We felt like we were on the
  241. 07:52right direction, but every time we kind
  242. 07:53of turned the conversation into, you
  243. 07:55know, what it would take for them to
  244. 07:57buy, there's just such a long list of
  245. 07:58things that the platform did not do that
  246. 08:01were really required of it. And here's
  247. 08:02where it's really tricky because,
  248. 08:04similar to my experiences at Neurop
  249. 08:05Plus, you can delude yourself into
  250. 08:06thinking you're just one feature away
  251. 08:08from product market fit all the time.
  252. 08:09And you can delude yourself into
  253. 08:10thinking like, I just got to keep
  254. 08:11building, keep building, keep building,
  255. 08:12and then one day I'll find it. And
  256. 08:14that's almost never the case. Now in our
  257. 08:16case we we had a theory that the product
  258. 08:19that we wanted to build had to be a
  259. 08:21platform. It had to combine the full
  260. 08:22ITSM this AI native workflow builder
  261. 08:26access management help desk automation
  262. 08:28that it was actually the full platform
  263. 08:31that was what was going to be successful
  264. 08:32in the market and we wouldn't really
  265. 08:34have a product to test until we built
  266. 08:36the platform. And so that's how we kind
  267. 08:39of kept the conviction during that
  268. 08:40one-year period where we're just
  269. 08:42building and we're not getting any kind
  270. 08:43of market traction is that well the the
  271. 08:46product is not built yet and we have to
  272. 08:47just build it and then we'll know
  273. 08:49eventually we close our first deal and
  274. 08:53then we close our second and we close
  275. 08:55our third and our fourth and then that
  276. 08:57just kind of creates this cascade and
  277. 08:59all these customers that are providing
  278. 09:00feedback and the product's getting
  279. 09:01better and we're iterating faster and
  280. 09:03then the product gets better and we
  281. 09:04close more deals and um every
  282. 09:06conversation ends up becoming
  283. 09:08not oh this is really interesting keep
  284. 09:10me posted but how much does it cost how
  285. 09:13do we try this out what's your PC
  286. 09:15process and so the the tenor of the
  287. 09:17conversation changed almost overnight
  288. 09:19over a few week period in April or May
  289. 09:21of 2025
  290. 09:23where it went from this is really neat
  291. 09:25keep us posted to how much does it cost
  292. 09:28how do we get started because in our
  293. 09:30early conversations we would get nuggets
  294. 09:32that were interesting of customer
  295. 09:34interest in what we were trying to build
  296. 09:36and we just didn't let that kind of like
  297. 09:38push or pull us. So, you become very
  298. 09:39neutral to the customer feedback.
  299. 09:40Whereas, I think in the early part of my
  300. 09:42career, every customer conversation, you
  301. 09:44know, you just you're a pendulum. And
  302. 09:45so, you have a good call and you're
  303. 09:46you're you're on the moon. They're so
  304. 09:48excited. Like, people love this idea.
  305. 09:50You have a bad call, you're just in the
  306. 09:52dumps. And in this company, when we were
  307. 09:54doing customer discovery, we just kind
  308. 09:57of took a lot of calls. We did not let
  309. 09:59any single conversation influence us too
  310. 10:01much. So I try to be on customer calls
  311. 10:04five or six hours every single day
  312. 10:05because it's not just an IT help desk
  313. 10:08automation platform. It is it is a broad
  314. 10:10enterprise automation platform for the
  315. 10:12organization. So I think uh a lot of
  316. 10:14folks approach these customer
  317. 10:16conversations as these point in time
  318. 10:18interviews, these kind of discreet
  319. 10:19moments where I'm going to do a customer
  320. 10:20interview, I'm going to do a customer
  321. 10:22conversation and I approach it much more
  322. 10:24as a relationship that is built and
  323. 10:27fostered over time. I want to be
  324. 10:28embedded with customers. So, I don't
  325. 10:31think about customer interviews or
  326. 10:32customer conversations. Like, I'm just
  327. 10:34talking to customers all the time. I'm
  328. 10:36in their Slack channels. I talk to them
  329. 10:37every single day. That builds just an
  330. 10:40intuition for our customers that is much
  331. 10:42much more powerful than these kind of
  332. 10:44point in time, hey, I've got a list of
  333. 10:46questions I want you to answer. You
  334. 10:47know, customer interviews are valuable,
  335. 10:49but I think what's more valuable is
  336. 10:51actually understanding the customer,
  337. 10:53understanding their problems, their
  338. 10:54pains, their perspective, how they use
  339. 10:57your product. And you can't build that
  340. 11:00understanding quickly. You can't build
  341. 11:02it through formal interviews. You can
  342. 11:04only build it by truly being embedded in
  343. 11:07their day-to-day and and building
  344. 11:08relationships with them over time. It
  345. 11:10has to be an a genuine interest in them
  346. 11:13and in making their lives better that
  347. 11:15drives that relationship forward and
  348. 11:17gives you the insights you need to build
  349. 11:19a better product.
  350. 11:21I think when you're building an AI
  351. 11:22product today, one of the most
  352. 11:24challenging things to always be thinking
  353. 11:26about is what's going to happen with the
  354. 11:28models and how are these capabilities
  355. 11:29going to change over time. And in the
  356. 11:31early days, to be honest, it didn't. You
  357. 11:32know, the workflow builder did not allow
  358. 11:34you to just describe any arbitrary
  359. 11:35automation and have that build out
  360. 11:37automatically without intervention. It
  361. 11:39it took a iteration. It took um
  362. 11:41prompting. It it took a lot of
  363. 11:42behind-the-scenes minations to make that
  364. 11:44all work and and be seamless. And I
  365. 11:46think the bet that we took was that
  366. 11:48between our efforts and improving the
  367. 11:50product and doing a lot of clever things
  368. 11:52under the hood and the improvements in
  369. 11:55the underlying models that we would be
  370. 11:57able to build a product that delivered
  371. 11:59on this vision of make automation. So
  372. 12:02I'm I'm very bullish that the software
  373. 12:04companies of the future are going to be
  374. 12:06much broader in scope, but I also think
  375. 12:07there are going to be more of them
  376. 12:09because people do want more out of these
  377. 12:12platforms. And so I don't I don't
  378. 12:13necessarily think there's going to be a
  379. 12:15consolidation. U but I do think that
  380. 12:18we're going to expect a lot more out of
  381. 12:19all these tools. Are you building for
  382. 12:22how the models work and behave today or
  383. 12:25are you building for a future where
  384. 12:27they're a little bit better or are you
  385. 12:29building for a future where they're a
  386. 12:30lot better and they work very very
  387. 12:31differently? And figuring out what is
  388. 12:33the right call is really hard. And on
  389. 12:36the one hand, you want to kind of be
  390. 12:38skating ahead a little bit and knowing,
  391. 12:41hey, like we did where it's like the the
  392. 12:43models are not good enough for what we
  393. 12:44want to do yet, but they will be very
  394. 12:46soon. In the same way, you can delude
  395. 12:47yourself into thinking, oh, this doesn't
  396. 12:50work today, but all we need is the
  397. 12:51models to get better and then all of a
  398. 12:52sudden we'll have a company. I think
  399. 12:54that's a bad idea. But I think it is
  400. 12:56smart to to embrace challenges that are
  401. 13:00almost possible and you could with a
  402. 13:04small small leap of faith believe that
  403. 13:06they're going to be possible in the near
  404. 13:08future.