Transcriptly

How We Captured Silicon Valley in just 4 months | Crush-it Conference, Rahul Vohra

EO15:34Added Aug 31, 2026

EO had a great opportunity to host the 'Crush-It Conference 2024' in San Francisco, featuring Rahul Vohra, Founder and CEO of Superhuman. In this session, we explore how Superhuman rapidly developed as a competitive, best-selling productivity AI tool in just 4 months. Superhuman AI pro

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. Intro

  2. 00:00I wanted to be able to get at it really, really quickly
  3. 00:02So it was small, it was fast, it was chaotic, it was stressful
  4. 00:06We did it really, really quickly
  5. 00:08And a few months later, I think it was June of last year
  6. 00:11We launched superhuman AI, the very first version, which lets you write with AI
  7. 00:15so you can jot down a few notes and we'll turn it into
  8. 00:18a fully written email that's in your own voice and tone
  9. 00:20We match all the emails you've already sent, this feature that we're now all
  10. 00:23very familiar with, and basically every application that we use,
  11. 00:25that was the very first thing that we built, and it really
  12. 00:28captured the hearts and minds of our user base of Silicon Valley
  13. 00:32It was it was a really big deal at the time
  14. 00:43My name is Rahul
  15. 00:43I'm the founder and CEO of superhuman, where we make the fastest
  16. 00:47email experience in the world
  17. 00:49With superhuman, you get through your inbox twice as fast as before,
  18. 00:53reply 1 to 2 days sooner and save four hours or more every single week
  19. 00:57All right, so Rahul, we're going to be talking about productivity tools today
  20. 01:01And you're building one
  21. 01:02And we're going to start off with the very first question while you were building

  22. Behind the Scenes of Developing Superhuman's AI

  23. 01:06superhuman AI. Can you share with us a little bit about when you first started?
  24. 01:09Okay, let's build this, and can you walk us through some of the steps
  25. 01:13that you took in order until it was born?
  26. 01:16What steps did you take to bring superhuman AI to the world
  27. 01:19and kind of challenges did you face?
  28. 01:21How did you incorporate the team?
  29. 01:22As much as you could tell, that would be super awesome
  30. 01:24I think we started last year in February and believe it or not, nobody really
  31. 01:29knew anything about AI. We didn't have a machine learning engineer on staff
  32. 01:32We'd never needed one
  33. 01:33It just didn't turn out to be important for what we were building
  34. 01:36We hadn't really used large language models ever
  35. 01:39We were sort of vaguely aware
  36. 01:40I mean, how could you not be aware of what was going on?
  37. 01:42But it was around January, February when it became incredibly clear to me,
  38. 01:46as I think it did all of us, that it's an existential moment
  39. 01:49And these moments don't happen particularly often
  40. 01:51in the history of startups,
  41. 01:52but it was obvious that it was existential that everything would change
  42. 01:56and that this is something that had to become one of the core pillars
  43. 01:59It's like a top three priority
  44. 02:00If we have three priorities, it's one of them at the company now,
  45. 02:04because we didn't really have anyone on staff who knew
  46. 02:08what they were doing, and we didn't have time to wait and hire for this
  47. 02:11And also it was something that was going to be existential to the company
  48. 02:13We actually approached it in a very different way
  49. 02:15than we do most new initiatives
  50. 02:18And so for context, we at the time, I think around 100 people,
  51. 02:21around 110 people
  52. 02:22Even so, we didn't assign this project a product manager, as you might add,
  53. 02:28a company of that size to run it
  54. 02:30We decided that I would actually run it personally,
  55. 02:32and fortunately I just hired a president
  56. 02:35So I was able to hand over day to day operation and management
  57. 02:38of the company to Paul our president
  58. 02:40And we created a small Tiger team, if you will, of myself,
  59. 02:44a designer and a handful of engineers
  60. 02:46And the idea was to essentially run it like a startup devoid of all process
  61. 02:50We just get together regularly and build stuff
  62. 02:52And with founder Fiat, I think a lot of what a founder has to do sometimes
  63. 02:56is in the absence of data, just say, this is what we're going to build
  64. 03:00Trust me, I will own it
  65. 03:01If it's wrong and if it's right, we'll move on and build the next thing
  66. 03:04But you need to be able to short circuits the decision paralysis
  67. 03:07That can sometimes happen, especially as companies get a little bit larger
  68. 03:10and debating things becomes the way in which decisions are taken
  69. 03:14And debating things is great
  70. 03:15That's actually how many decisions should be taken
  71. 03:18But for something brand new like this,
  72. 03:21I wanted to be able to get at it really, really quickly
  73. 03:23So it was small, it was fast, it was chaotic, it was stressful
  74. 03:27We did it really, really quickly
  75. 03:29And a few months later, I think it was June of last year

  76. Launched Superhuman AI in just 4 months

  77. 03:32We launched superhuman AI, the very first version, which lets you write with AI
  78. 03:36so you can jot down a few notes and we'll turn it into a fully written email
  79. 03:39that's in your own voice and tone
  80. 03:41We match all the emails you've already sent, this feature that
  81. 03:43we're now all very familiar with, and basically every application that we use,
  82. 03:46that was the very first thing that we built, and it really
  83. 03:48captured the hearts and minds of our user base of Silicon Valley
  84. 03:53It was it was a really big deal at the time,
  85. 03:56and that's an example of what we call on demand AI
  86. 03:59Relatively easy to run, cheap to run as well because it's only operating
  87. 04:03when people actually click the button or when they use the feature
  88. 04:05But that gave us the confidence to move on to the next phase of AI,
  89. 04:08which we call always on AI. So you move from on demand to always on,
  90. 04:11and this is AI that is just constantly running in the background all the time
  91. 04:15Much harder to build, much more expensive to run
  92. 04:18And the very first example of that we built was auto summarize
  93. 04:21So imagine when you receive an email, a one line summary
  94. 04:24that is pre-computed and always available on top of that conversation
  95. 04:28And when new emails arrive, it instantly updates
  96. 04:30And if you want to find out more, you just hit a shortcut and you can see
  97. 04:33some bullet points about the conversation
  98. 04:35People really love that
  99. 04:36And that then gave us the confidence to build Instant Reply,
  100. 04:39which is also an always on feature
  101. 04:41Imagine waking up to an inbox where every single email already has a draft reply
  102. 04:45You would just hit edit and then send, and sometimes you wouldn't even need to edit
  103. 04:48And that then gave us the confidence to build the next thing
  104. 04:50So we've been increasingly working up this confidence ladder,
  105. 04:56I would say taking on more ambitious, more expensive projects
  106. 04:58And now I'm not an ICI on that team
  107. 05:00But for the first six months I was and it was really to kickstart it
  108. 05:03Now it's being run like a regular project at superhuman
  109. 05:06Well, this is one of the great things about having a great leadership team
  110. 05:10And as I mentioned, I just hired a president
  111. 05:12So I was able to leave the explanation and the delegation
  112. 05:17and the authorization of this idea to Paul
  113. 05:18So he explained to everyone what was going on
  114. 05:21He figured out who the resourcing
  115. 05:22I didn't really have to explain or sort of
  116. 05:26go through any difficult conversations to make this happen
  117. 05:27Now, all these other AI agents that are coming up
  118. 05:30that doesn't just deal with email,
  119. 05:32there are these everybody is building all these AI tools to help you answer emails,
  120. 05:38book a call, book calendars for you
  121. 05:40And there are also other email softwares that are trying to emulate
  122. 05:44a lot of the features that have been popular in superhuman as well
  123. 05:47How do you see all this AI agents coming along
  124. 05:51and all this competition coming along
  125. 05:52And with the AI, how are you seeing this landscape
  126. 05:56and how do you see superhuman sort of moving the next steps forward?

  127. Don’t Obsess Over the Competition

  128. 05:59So this topic of competition has obviously come up so many times
  129. 06:03over the years at superhuman
  130. 06:04Every single time there's been a new email client, for example, in the early days,
  131. 06:08maybe some people in the company would panic and I would talk about it and,
  132. 06:12you know, we'd get on with building great software
  133. 06:14and delivering value to our customers
  134. 06:16And every single time it's just come and gone and come and gone and come and gone
  135. 06:19And Paul Graham has a good piece of advice about this,
  136. 06:23which is don't obsess over the competition
  137. 06:24It actually doesn't really matter
  138. 06:26Just focus on making your own customers, your own users, incredibly happy
  139. 06:31So that's what we've always done
  140. 06:32And I think the the AI wave is no different
  141. 06:34I think the Asian wave is no different
  142. 06:36We see superhuman as a core part of this evolution of the productivity landscape
  143. 06:41We are working really hard towards building AI Agentic future
  144. 06:45We believe in a world where one day you will have a superhuman AI agent,
  145. 06:49and you can start to see if you use the product today
  146. 06:52The beginnings of that with features like ask AI
  147. 06:55You can imagine an AI that can triage your inbox on your behalf,
  148. 06:58that can schedule meetings on your behalf, that can write,
  149. 07:02and one day even send fully written emails on your behalf
  150. 07:03An AI agent that can work not just on tasks, but complex problems
  151. 07:07Break them down into subproblems and work on them
  152. 07:11in parallel or in SQL that can resolve ambiguity
  153. 07:13Ask you clarifying questions
  154. 07:14Go off and research things, interrogate internal systems,
  155. 07:18APIs, or even other agents
  156. 07:20And I think that world isn't tomorrow's world
  157. 07:22Obviously, I've just described a lot, but over the next year or two,
  158. 07:25we will actually get there in some places faster than in others
  159. 07:28That is what we're building towards
  160. 07:29So let's get into some other questions,

  161. How to Make Customers Willingly Pay for the Product

  162. 07:32which is about building a productivity software
  163. 07:34Superhuman
  164. 07:34When people say they love productivity tools, there are many
  165. 07:38But then when it comes to paying for productivity tools,
  166. 07:41I feel that people shy away from it 'Oh yeah, I love productivity'
  167. 07:44And 'oh, it's paid' and kind of don't use it
  168. 07:46But for superhuman, from the very early on,
  169. 07:47you've had people line up to use your very first product
  170. 07:51What do you think makes that difference about a product
  171. 07:53that people are willing to pay for that productivity?
  172. 07:56What edge did you have at the very beginning that made it successful for you?
  173. 07:59And in terms of why folks have and are and will continue to be willing to pay,
  174. 08:04let's actually think about who are the people who send and receive
  175. 08:08a ton of email, who are the people who do email for a living?
  176. 08:12So founders, CEOs, obviously, most people in this room, if you're not
  177. 08:16already doing email for a living as your companies grow, you soon will be
  178. 08:19But it goes much deeper than that
  179. 08:20It's also account executives, account managers, customer success, business
  180. 08:25development, talent recruiting, business owners, realtors, leaders,
  181. 08:30and managers of every kind that you can imagine
  182. 08:32They all do email for a living, and a core part of their jobs
  183. 08:36actually is to be responsive, because if they're not responsive,
  184. 08:39then they can block their team, miss opportunities or in the worst case,
  185. 08:43even damage their own reputations
  186. 08:45So there's both a carrot and a stick there
  187. 08:47There's the danger of doing email poorly, which sadly, most people do
  188. 08:51And then there's also the benefit of doing it really, really well
  189. 08:53because it's clear to the job
  190. 08:55That's why people are willing to pay. So how do we define productivity?

  191. How to Define and Measure 'Productivity'

  192. 08:58I think most companies start with, and we certainly started this way
  193. 09:02with the time that we're saving people
  194. 09:05And there's two ways to do that. First of all, you can measure it directly
  195. 09:08You can measure how long it takes to do a certain activity inside of superhuman
  196. 09:12versus, for example, how long it takes to do those activities in Gmail or Outlook
  197. 09:15Measure how many of those activities people do and multiply it all out,
  198. 09:18and you'll get that kind of bottoms up measure
  199. 09:20But you can also do it top down as well, and we do both top down
  200. 09:23You would actually survey your users at superhuman
  201. 09:25We're pretty famous for our product market Fit Engine, which is a completely different
  202. 09:29kind of survey that is designed to get you to product market fit
  203. 09:32As part of that, we actually ask people, how much time are we saving you?
  204. 09:36And we get responses back so we can triangulate time saved
  205. 09:39both bottoms up and top down
  206. 09:40And that's how we get to that four hours or more every single week
  207. 09:44Answer. But that's just the beginning
  208. 09:45There's lots of other things that you care about
  209. 09:47when you're measuring email productivity
  210. 09:49So one example is how much email you actually get through
  211. 09:52One of the things that's kind of unique about being a founder
  212. 09:55or one of these high email roles, is there is an infinite number of emails to answer
  213. 10:00I mean, you may not actually be able to get through all of them
  214. 10:02So for some of us, inbox zero, at least in total over
  215. 10:07the entirety of the inbox may actually just be impossible
  216. 10:09So how do we start to measure productivity?
  217. 10:11Another way of doing it is the rates at which you respond to emails
  218. 10:14So let's say you work for an hour
  219. 10:15Are you responding to ten, 50, 100, 200 or more?
  220. 10:19And with superhuman, what we're able to show is that people in the same amount of
  221. 10:23time respond to 2.4 x 2.4 times the number of emails
  222. 10:29they would compared to in Gmail. So that's the second metric
  223. 10:30The third would be how rapidly you're responding to emails
  224. 10:33So folks may have seen this interview with

  225. The Big Difference Between Great Founders and Bad Founders

  226. 10:37Sam Altman back when he was the president of Y Combinator
  227. 10:38He noticed that there is a big difference between excellent founders
  228. 10:43and bad founders in terms of how they respond to emails
  229. 10:45And he defined excellent founders as those that went on
  230. 10:47to build billion dollar companies
  231. 10:48Bad founders as those whose companies eventually went on to fail, unfortunately,
  232. 10:52and obviously, as you would expect, the bad founders days
  233. 10:56or weeks to respond to an email
  234. 10:57Good founders, they responded within minutes, literally minutes
  235. 11:01The billion dollar company founders that you know today with the product,
  236. 11:04then without the product
  237. 11:06Can you tell us a bit more about how you view email as not just

  238. Email as a Collaboration Tool

  239. 11:09a communication tool, but something more?
  240. 11:11- If you do, we think. - It's not just communication
  241. 11:13We think it's also fundamentally a collaboration tool
  242. 11:17And we believe that so much.
  243. 11:18We actually reimagined the entire email experience to be multiplayer.
  244. 11:23So now inside of superhuman, what you can do is you can take any email
  245. 11:26and you can share a live view of that email with your team
  246. 11:30You no longer have to repeatedly forward or ask or take screenshots
  247. 11:34And as new emails arrive on that conversation, your team can see
  248. 11:37what those new emails are, even if they're not an email recipient directly,
  249. 11:41they can see that either in superhuman or they can go to a web page,
  250. 11:44and then you and your team can comment directly on those emails
  251. 11:47And those comments aren't email objects.
  252. 11:50There's no risk that they leak, They're stored entirely separately
  253. 11:51And you can have a safe conversation about whatever it is you're trying to do
  254. 11:55So we believe in this collaborative future of email so much
  255. 11:59that we actually rethought the entire experience to build these features
  256. 12:02And it was a huge lift I mean, it took us
  257. 12:03well over a year to do so, but that's the kind of thing
  258. 12:06we've been building
  259. 12:07Can you talk a bit more about what change
  260. 12:07it has been for you to build email as a collaboration tool
  261. 12:13and offer it to teams on what's so different about it,
  262. 12:16and what makes it better for your company too
  263. 12:19just over the course of the last year, have scaled from relatively small and mid-market startups
  264. 12:24to teams of several thousand, which is really exciting to see
  265. 12:27And that kind of revenue ramps really quickly
  266. 12:29Where I would have gone multiplayer sooner is it's much easier to build a strong
  267. 12:34business when you're selling to teams
  268. 12:36When you sell to an individual. It's a double edged sword
  269. 12:39The good news is that you only need to convince one person to adopt the tool,
  270. 12:44and that means you can tell a very clear and coherent story about,
  271. 12:47Hey, you do a lot of email
  272. 12:48You're the founder, you're the CEO, or you're the salesperson
  273. 12:51Imagine being able to do it twice as fast, or respond more quickly,
  274. 12:54or being able to work more deals or close deals faster
  275. 12:56It lets you tell a very clear and pointed story
  276. 13:00However, ultimately it's harder for that person to stick around because at some point
  277. 13:04someone else may come along and tell an equally clear and as pointed story
  278. 13:08And in general, there have been very, very few companies in software that have
  279. 13:12scaled to be extremely large off the back of single player subscriptions alone,
  280. 13:18and we could probably name them right here
  281. 13:20They're companies like Spotify and Netflix
  282. 13:23and trying to think what else it might be
  283. 13:25Very, very few that have scaled to truly billion dollar
  284. 13:28single player subscription companies,
  285. 13:30and the things that make those companies feasible
  286. 13:34is that they can credibly and reliably, year after year
  287. 13:36Add tens, if not hundreds of millions of new subscribers
  288. 13:40to the top of the funnel every single year
  289. 13:41Because if you look at the retention numbers, they're only
  290. 13:43retaining 50 to 60% of those subscribers at the end of any given year
  291. 13:48And when they stop being able to do that, they just have to start raising prices
  292. 13:51And this is what Netflix does
  293. 13:52And turns out not to be a particularly great business in the very long haul,
  294. 13:55unless you can continuously add those users
  295. 13:58So if you're building a subscription business,
  296. 14:00it is a much more robust to build something that you can sell to teams
  297. 14:06It's harder in other ways, because now you have to convince
  298. 14:08a team of people to adopt it
  299. 14:09But once a team of people does adopt it, it is so much stickier
  300. 14:13It is so much harder for them to churn out and to use something else
  301. 14:16And at that point, you're able to get to what a lot
  302. 14:18of investors call negative churn
  303. 14:21This is when, yes, you're going to lose some seats in a company
  304. 14:24or even in a cohort of companies,
  305. 14:26but the seats that you acquire during that period in that same cohort of companies
  306. 14:31will more than offset those
  307. 14:33seats that you lose such that any particular cohort
  308. 14:36is going to be expanding over time in terms of revenue,
  309. 14:39even if certain accounts perhaps churn out to zero
  310. 14:42And that is actually what most SaaS B2B businesses do
  311. 14:45They grow off the back of negative churn off those cohorts expanding
  312. 14:49So we did the long, hard grind of building a single player, first of making it work
  313. 14:53with Gmail, with office 365, on iPhone, on Android, on Mac, on Windows and Web
  314. 14:58And that took many, many, many years
  315. 14:59We did that and made it something that individuals actually wanted to use
  316. 15:03before finally coming back and making it truly multiplayer
  317. 15:05over the course of the last year,
  318. 15:09I didn't want to build an email client that your boss forces you to use
  319. 15:11I wanted to build a piece of software that you genuinely,
  320. 15:14really fell in love with and that you wanted to use
  321. 15:16All right, well, thank you everyone.Thank you, Robin, for the amazing talk
  322. 15:20Let's give him a round of applause