How a Cockroach Startup Became a $2.2B Market Leader | EliseAI, Minna Song

EO15:00Added Aug 31, 2026

Minna Song, co-founder & CEO of EliseAI turned 100 VC rejections into a $2.2B company after 8 years of relentless hustle and survival. While others chased hy...

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format
  1. Intro

  2. 00:00If you want to be the best company in
  3. 00:01the world, I think you have to work the
  4. 00:02hardest in the world. It's like anything
  5. 00:04else. If you want to be the best
  6. 00:05athlete, you want to be the best
  7. 00:06musician, you're practicing non-stop.
  8. 00:08Work in a tech startup is exactly the
  9. 00:10same thing. When you don't have a lot of
  10. 00:12money, that's not something that you can
  11. 00:14kind of fall back on. You have to do
  12. 00:16everything yourself. And you should talk
  13. 00:18to all of your customers, handle all of
  14. 00:20your customer success, support tickets.
  15. 00:22You should do sales yourself. that helps
  16. 00:25you really focus on exactly what your
  17. 00:27customers are asking for and what
  18. 00:29they're willing to pay for. When we
  19. 00:31launched our product with our first
  20. 00:33national customer, we were just two
  21. 00:35people at the time. It was just me and
  22. 00:37and Tony, my co-founder. It was a
  23. 00:3924-hour service. So, that meant he
  24. 00:42stayed up until 4:00 a.m. and I woke up
  25. 00:44at 4:00 a.m. and watched it. That was
  26. 00:47probably the most physically draining
  27. 00:49time, but we were able to survive. Yeah,
  28. 00:52it was a really exhausting time of
  29. 00:54course, but it doesn't really matter.
  30. 00:56You have problems. You have customers
  31. 00:59that need things. It's just about every
  32. 01:01single day doing everything that you can
  33. 01:04prevent your company from dying. And
  34. 01:06even when you're exhausted, I think
  35. 01:07that's a pretty motivating. It's a
  36. 01:09stressful thing to start a company. It's
  37. 01:10a stressful thing to work at a at a
  38. 01:13startup that's growing. You have to take
  39. 01:15one step in front of the other just to
  40. 01:17solve problems over time. being a
  41. 01:18cockroach where you can't let anything
  42. 01:21kill you ever.
  43. 01:24My name is Minong and I'm the co-founder
  44. 01:26and the CEO of Elise AI. Elise AI is a
  45. 01:29tech company that uses AI to drive real
  46. 01:32world impact and we are starting with
  47. 01:33two of the most critical industries that
  48. 01:35make up over 40% of the US GDP. the
  49. 01:38housing industry and the healthcare
  50. 01:39industry. Elise AAI is used by over 10%
  51. 01:43of the US apartments today and over
  52. 01:45twothirds of the largest property
  53. 01:47management companies use our software.
  54. 01:49We raised a $250 million series E round
  55. 01:52which was led by Andre and Horowitz and
  56. 01:55participated in by Bessemer Venture
  57. 01:56Partners as well as previous investors
  58. 01:58Sapphire Ventures and Navatas Capital.
  59. 02:00Our company has been doubling
  60. 02:02consistently year-over-year for the last
  61. 02:04four years, and we passed 100 million
  62. 02:06annual recurring revenue earlier this
  63. 02:08year.
  64. 02:10[Music]
  65. Why 90% of Founders Fail to Find PMF (And How Not to Be One of Them)

  66. 02:18I met my co-founder in college. So,
  67. 02:21while I went to MIT, he actually met him
  68. 02:24studying abroad at the University of
  69. 02:25Cambridge in the UK. So we we both come
  70. 02:28from software engineering backgrounds
  71. 02:29and felt that we should apply technology
  72. 02:32to doing something good in the world. I
  73. 02:34think it's pretty obvious tech is the
  74. 02:36fastest way to bring massive scale
  75. 02:39impact in the world. So it was pretty
  76. 02:41clear that we were going to start a
  77. 02:42technology company someday. When we
  78. 02:44started Elise, we actually wanted to
  79. 02:47start understanding the industry that we
  80. 02:49were operating in and really knew
  81. 02:51nothing about the housing industry
  82. 02:53except that we wanted to do something
  83. 02:55really, really big in it. And so when I
  84. 02:57got started, I started working at a real
  85. 03:00estate firm here in New York City. And
  86. 03:02that gave me a great opportunity to meet
  87. 03:04a lot of people. So I took a front desk
  88. 03:06job where I greeted everyone who came
  89. 03:09through the door. I picked up all the
  90. 03:11phone calls and that's how we got the
  91. 03:14experience and understanding of the
  92. 03:15industry that helped us decide what we
  93. 03:17wanted to build. I uh actually don't
  94. 03:19think it's extreme at all. I think it's
  95. 03:22actually quite insane when people don't
  96. 03:24try to spend time really understanding a
  97. 03:26problem before they try to solve it and
  98. 03:28then they waste years and years of
  99. 03:30building technology that never finds
  100. 03:32product market fit. So, I think this was
  101. 03:34maybe one of the most obvious uh and
  102. 03:37rational decisions that we've ever made,
  103. 03:39which is just just learn learn about the
  104. 03:41people that you're building for. Learn
  105. 03:43about the problems. Uh especially
  106. 03:46because we didn't come from a background
  107. 03:47in housing at all. It's really
  108. 03:49important, especially in industries like
  109. 03:51housing and healthcare where so much
  110. 03:53happens in person. It doesn't all happen
  111. 03:56in the digital world. So, you can't just
  112. 03:57look at metrics. You can't look at users
  113. 04:00and how they interact with your online
  114. 04:02products. You have to be there to
  115. 04:04understand how they're interacting with
  116. 04:06the real world. The problem was pretty
  117. 04:08obvious. I would say that it wasn't
  118. 04:10something that we had an aha moment. It
  119. 04:13was just something people were
  120. 04:14constantly complaining about. We just
  121. 04:16keep tackling those one by one. Their
  122. 04:18frustrations, they couldn't reach a
  123. 04:20building. No one's answering their phone
  124. 04:21calls. No one's answering their emails.
  125. 04:23And so that's the exact problem that we
  126. 04:25tried to solve. It wasn't really a
  127. 04:26question about what the problem was. It
  128. 04:28was really a question of is the
  129. 04:31technology the right technology to solve
  130. 04:33the problem and was it the right time in
  131. 04:362017? A lot of people didn't know what
  132. 04:38AI was. A lot of people didn't believe
  133. 04:41that AI could answer their emails. A lot
  134. 04:45of people didn't believe that AI would
  135. 04:47do things better than humans in any
  136. 04:49capacity. A lot of people thought this
  137. 04:51market was small because companies that
  138. 04:53had been built in the industry before
  139. 04:56never grew to be that large of
  140. 04:58companies, but I can see that other
  141. 05:00people over the years have have not seen
  142. 05:03real estate as a sexy industry. But I
  143. 05:06think housing was always sexy to us
  144. 05:08because it's a fundamental human need.
  145. 05:10It's an incredibly important part of
  146. 05:12society and our lives. So why wouldn't
  147. 05:15that be sexy?
  148. Brutal Lessons From Our First 10 Customers

  149. 05:20The first 10 customers were really
  150. 05:21brutal. We were cold calling. We were
  151. 05:25knocking on doors all across New York
  152. 05:26City. We got lucky with a few customers
  153. 05:29who signed on and then we started
  154. 05:31selling to national property management
  155. 05:33companies. It was pretty crazy that we
  156. 05:35ended up signing the first and second
  157. 05:37largest apartment owners at the time
  158. 05:39onto our platform for their entire
  159. 05:41portfolio. And we were just two people
  160. 05:44at the time. It was just me and and
  161. 05:46Tony, my co-founder. So, we had these
  162. 05:49large customers buying into a product
  163. 05:51that two people had built that was meant
  164. 05:54to run their entire real estate
  165. 05:56portfolio.
  166. 05:58How we convinced these large customers
  167. 06:00to work with us. I remember I went to a
  168. 06:03meeting. It was me and 12 executives, so
  169. 06:07all of their sea level team essentially
  170. 06:09sitting in a room with just me. I was in
  171. 06:12my young 20s. It was three hours of
  172. 06:14asking me a ton of questions. Can your
  173. 06:16product do this? Can your product do X?
  174. 06:18Can your product do Y? And honestly, I
  175. 06:20think I said for three hours straight, I
  176. 06:22just said, "No, it doesn't do that." And
  177. 06:24um I walked out of that meeting, called
  178. 06:26my co-founder and said, "This was the
  179. 06:28worst meeting of our lives. Just forget
  180. 06:30it. That deal is dead." And then the
  181. 06:32next day they called me back. The COO
  182. 06:34called me and said, "We're going to do
  183. 06:36this together. We're going to help you
  184. 06:37build the product." They knew we were
  185. 06:39early. We were not hiding the fact that
  186. 06:41we were a tiny startup, but they
  187. 06:43believed in us because they believed
  188. 06:45they had a big problem and that they
  189. 06:47could work with honest founders to
  190. 06:49actually solve that. I think when you're
  191. 06:51working with your first few customers,
  192. 06:52you have to build a lot of things that
  193. 06:54they ask for. A lot of people are really
  194. 06:56afraid of building too much or
  195. 06:57overfitting for certain customers, but
  196. 06:59the reality is that fear will paralyze
  197. 07:02you. You really just need to build what
  198. 07:03they ask for and then as you get
  199. 07:05customer two, three, four, it'll give
  200. 07:07you a lot more data points. And maybe
  201. 07:09you've built something that's not
  202. 07:11sustainable in the long run, but quite
  203. 07:13honestly, I believe that your first
  204. 07:15customers, you will have to service that
  205. 07:18forever. That's a really common
  206. 07:20attribute of early adopters is they want
  207. 07:22a lot of control. That's why they picked
  208. 07:24to work with an early stage startup in
  209. 07:26the first place. But they do want a lot
  210. 07:28of control and usually they want to
  211. 07:30dictate what your road map looks like.
  212. 07:32I'd say don't be afraid. I I think you
  213. 07:34will make mistakes. But that's a normal
  214. 07:36process I think in in building a massive
  215. 07:39company.
  216. Only Cockroach Founders Survive

  217. 07:43In our first round of funding, we had
  218. 07:45literally over 100 rejections. A lot of
  219. 07:47the rejections were the same. They're
  220. 07:49like, "We're not interested in this
  221. 07:51market. We only look at horizontal SAS
  222. 07:54or we only look at enterprise SAS and
  223. 07:58things in the housing industry were not
  224. 08:00really interesting to them. When we
  225. 08:01didn't have a lot of funding, we really
  226. 08:03just relied on ourselves and our ability
  227. 08:05to build features that our customers
  228. 08:07were asking for. I think we were cash
  229. 08:10constrained. So, we didn't have huge
  230. 08:11engineering team to build lots of
  231. 08:13features in parallel. But I think that
  232. 08:15prioritization helped us build the
  233. 08:17things that were absolutely the most
  234. 08:19critical that had the highest amount of
  235. 08:21value, not ancillary features or
  236. 08:24features that were just providing kind
  237. 08:25of optics for customers to to think that
  238. 08:29they were kind of featureful products.
  239. 08:31So I think money doesn't actually solve
  240. 08:33a lot of problems. I think when people
  241. 08:35raise a lot of funding early on, they
  242. 08:37they think that they're successful and
  243. 08:39they throw a lot of money at problems
  244. 08:41and and think that those problems are
  245. 08:42solved. But really what happens is those
  246. 08:45problems end up doing and growing and
  247. 08:48and they're still there. And when you
  248. 08:50don't have a lot of money, that's not
  249. 08:52something that you can kind of fall back
  250. 08:54on. Your problems are very obvious and
  251. 08:57you know if you solve them or if you
  252. 09:00hadn't solved them. Eventually, as we
  253. 09:02started building a track record of
  254. 09:04showing real growth in the company,
  255. 09:06investors started to see what was
  256. 09:08possible. Yeah, it was a really
  257. 09:10exhausting time of course, but it
  258. 09:12doesn't really matter. You have
  259. 09:14problems. You have customers that need
  260. 09:16things. You have to take one step in
  261. 09:18front of the other just to solve
  262. 09:19problems over time. Being a cockroach
  263. 09:21where um you can't let anything kill you
  264. 09:24ever.
  265. The Secret to Hypergrowth

  266. 09:28What's our secret sauce? I think a lot
  267. 09:30of our growth has come from building a
  268. 09:31lot of products. No matter what customer
  269. 09:34we're pitching to in the housing
  270. 09:36industry, we have a product that solves
  271. 09:38the problem that they are thinking about
  272. 09:40today. So whether that's something about
  273. 09:41their leasing process, whether it's
  274. 09:43something about their maintenance
  275. 09:44process, we have a solution for them and
  276. 09:46that's really just come from having
  277. 09:48really high product velocity as a
  278. 09:49company. Another big reason is that
  279. 09:51we've just hired really, really well.
  280. 09:52I've spent most of my time focused on
  281. 09:54getting the best and most talented
  282. 09:56people I can possibly get. I
  283. 09:58interviewed, personally, interviewed the
  284. 10:00first 400 hires that we made. It's been
  285. 10:02incredibly important to have really
  286. 10:04strong people because when you're moving
  287. 10:06so fast and you have so many different
  288. 10:08products, it turns out that a lot of
  289. 10:10people have to make decisions
  290. 10:12themselves. And you need to have really
  291. 10:14smart and really hardworking people that
  292. 10:15are going to move the company forward.
  293. 10:17Because if you're moving quickly, the
  294. 10:19CEO cannot make all the decisions for
  295. 10:21you. The CTO cannot make all of the tech
  296. 10:23decisions for you. You need everyone to
  297. 10:25have really strong business acumen. One
  298. 10:27of the most painful hiring mistakes I
  299. 10:29made was probably around 2021 when we
  300. 10:33were raising our first big round of
  301. 10:35funding, our series B. And um I hired a
  302. 10:37ton of people who were really good at
  303. 10:40specialist roles. And the challenge I
  304. 10:42was trying to solve was that everyone
  305. 10:44everything was so chaotic. Every early
  306. 10:47stage employee was working on five
  307. 10:49different things and the constant
  308. 10:51complaint was I can't get anything done
  309. 10:54well because I keep switching context
  310. 10:57all the time which is a very common
  311. 10:58early stage company problem. And then I
  312. 11:01decided to solve that by hiring people
  313. 11:02who were kind of really good
  314. 11:04specialists. So if somebody was doing
  315. 11:06five things now they could pick the one
  316. 11:09that they really wanted to do and then
  317. 11:10you'd hire four other people that were
  318. 11:12specialists. The problem is everyone
  319. 11:14started losing the context. So they'd be
  320. 11:17building really well but in five
  321. 11:19different directions and the company was
  322. 11:21not moving forward. We were not building
  323. 11:22products that our customers cared about.
  324. 11:24We were not selling products that our
  325. 11:26customers cared about and it was just
  326. 11:28way too early to make those types of
  327. 11:30hiring decisions. So, I went back to
  328. 11:32really valuing generalist hires and I
  329. 11:36found that we could move a lot quicker
  330. 11:39if a lot of people were able to context
  331. 11:42switch were able to have success across
  332. 11:45many different domains, whether that was
  333. 11:47sales or customer success. They could
  334. 11:50think technically at the same time as
  335. 11:51speaking with customers. And those ended
  336. 11:54up being some of our most valuable hires
  337. 11:55for a very long time.
  338. Forget Shortcut, Just 24/7 Grind

  339. 12:00When we launched our product with our
  340. 12:03first national customer, we needed to
  341. 12:05keep the service running 24/7 and this
  342. 12:08was the most important deal of our
  343. 12:10lives. It had to go perfectly. So Tony
  344. 12:13and I were monitoring the system 24/7
  345. 12:15because we knew this was our shot. He
  346. 12:17stayed up until 4:00 a.m. and I woke up
  347. 12:20at 4:00 a.m. and watched it. And by the
  348. 12:22way, at the same time, we were building
  349. 12:24new features with this customer. We were
  350. 12:26trying to sell new customers. We were
  351. 12:29trying to support our existing customers
  352. 12:31and other portfolios that we had signed
  353. 12:33on. So that was probably the most
  354. 12:35physically draining time. I still work 7
  355. 12:38days a week 90 hour 100 hours is not
  356. 12:40really sustainable. That's when you're
  357. 12:42actually starting to eat into sleep and
  358. 12:44and deteriorating over time. But 90
  359. 12:47hours is sustainable. It just means you
  360. 12:49wake up, you go to work and you go home
  361. 12:51and you go to sleep and you get proper
  362. 12:53sleep and then you just do it again
  363. 12:54tomorrow. And you do that every single
  364. 12:56day. And that is how problems get
  365. 12:58solved. I think that my suggestion for
  366. 13:00founders is always if they're not
  367. 13:02working 7 days a week, they're not
  368. 13:04growing their company as fast as they
  369. 13:05should. Their company is going to be
  370. 13:07beaten by another company that is
  371. 13:09working that their founders are working
  372. 13:117 days a week. So that's just extra time
  373. 13:14and effort that will accelerate the
  374. 13:16growth of your company. If you want to
  375. 13:18be the best company in the world, I
  376. 13:19think you have to work the hardest in
  377. 13:20the world. It's like anything else. If
  378. 13:23you want to be the best athlete, you
  379. 13:24want to be the best musician, you're
  380. 13:26practicing non-stop. Tech startup is
  381. 13:27exactly the same thing. The culture here
  382. 13:29is filled with a lot of people who just
  383. 13:31really like solving hard problems. And
  384. 13:33you're surrounded by so many people that
  385. 13:35have this type of culture. There's this
  386. 13:37kind of intense gravity that when you
  387. 13:40bring in new people, they feel that
  388. 13:42energy and they really want to meet that
  389. 13:44bar.
  390. 13:46I actually think that most of the next
  391. 13:48unicorns will be products that are built
  392. 13:51in unsexy industries. I think a lot of
  393. 13:53the most important problems that are
  394. 13:56left to solve in the world are problems
  395. 13:58that are typically deemed unsexy. If
  396. 14:00they seem sexy, that means people are
  397. 14:03already very excited about them and
  398. 14:06they're already trying to solve them.
  399. 14:07And so long term, everything that we try
  400. 14:10to do in housing and healthcare is to
  401. 14:13make these industries much more
  402. 14:15financially stable and much more
  403. 14:17attractive so that we can have more
  404. 14:18housing providers and we can have more
  405. 14:20healthcare providers and that serves all
  406. 14:22of us. And we want capital to be put in
  407. 14:25the most fundamental and important parts
  408. 14:27of society that serve us as humans.