Rahul Vohra: Building AI Tools That Save Us Time

0:28Hey, everyone, and welcome to Generative Now. Coda and Placer. We talk about his philosophy as an investor and founder, his views on where agentic AI is headed, and what's next for Superhuman. So let's dive in. Rahul, so good to see you. Good to see you. Thanks for having me. Thanks for being on. So much I want to talk about with you. Obviously, there's recently been some big news about Superhuman. You're joining Grammarly, which is super exciting. I want to talk all about that. But just for the sake of anyone who's listening or watching who maybe doesn't know about superhuman, which I would find nearly impossible.

1:04They must be living under a rock if they don't. Tell us a little bit about superhuman, what it is, how you got started, and we'll go from there. Well, for folks who haven't heard, superhuman is the most productive email app ever made. Imagine getting through your email twice as fast as before, responding faster to the things that matter, and saving four hours or more every single week. On top of that, imagine waking up to an inbox where every email already has a draft reply. You would simply edit and then send. And sometimes you wouldn't even have to edit. It's really an amazing product. And I believe, at least for me, and I know for so many other people I work with or know of personally, like it feels like superhuman has kind of changed their lives.

1:45It takes this really, really daunting task of getting through your email, which seems to be getting harder constantly with the amount of spam and cold outreach we all get. actually manageable and in many cases delightful. There's so much unpacked with superhuman. First of all, you know, I was always blown away by the speed of the product. And then second of all, another thing that always stuck out to me about the product was sort of like the distribution strategy and the pricing. And if I remember correctly, and maybe it's still this way, I'm not even sure now, it's sort of like always in a state of beta.

2:19It almost felt like Like you're in this sort of exclusive club if you're using Superhuman. You're getting this test flight, but you're also paying for it. Like give us some of the insight around some of these key breakthroughs with Superhuman. Sure. Starting with the first piece, when I started Superhuman, I wanted to solve the biggest possible problem. And email is actually a way bigger problem than most people realize. I think it's one of those Silicon Valley blind spots. We're actually predisposed in Silicon Valley to, because so much of us use Slack every day to not realize how big of a problem email is.

2:52And it hit me just after I'd sold my last company, Reportive, which was also an email company, to LinkedIn. And I was commuting from work in Mountain View to where I lived in San Francisco. And for any of us who had the dubious pleasure of that commute, it's a one-hour journey. It's quite a while. But this was in 2014, and both Uber and Lyft were having their moment. So I happened to be in the back of an Uber. and on this trip I finished writing some product documents, clearing out my email, checking my calendar for tomorrow, making dinner plans with my partner. I even called my mother. So I was simultaneously being a productive employee, a proactive boyfriend, and a present son.

3:31And I realized in that moment that the real magic of Uber was not commuting. It was not about getting from A to B. It was time, being able to do all these other things. So I asked myself the question, where do we spend all of our time? And it turns out that there are roughly 1 billion professionals on the world. And on average, we spend three hours a day reading and writing email. That's 3 billion hours every single day or north of a trillion hours every single year. In fact, I couldn't find anything that we spend more time on. We all spend more time on email than we do any other aspect of our work.

4:10Now at the same time I could see Gmail and Outlook getting worse every single year becoming more cluttered using more memory consuming more CPU slowing down your machine still not working properly offline and on top of those products people were installing plugins like my last company Reportive but you know you'll also know of companies like Boomerang, Mixmax, Clearbit, you name it people had it and yet each plugin took those problems I just mentioned of clutter, memory, CPU performance offline and made all of them dramatically worse. So we decided it is time for change. We set out to rebuild the entire email experience from the ground up.

4:46And like I said, today, our customers are getting through their email twice as fast as before, responding faster to the things that matter and saving four hours or more every week. It sounds like time is basically integral to the offering. And if time is integral to the offering, you've got to make the thing as fast as humanly possible. I'm sure there's a lot more to it, But I'm guessing it starts with that. It starts with sort of the mission that you already laid out. It does. But it's not just the raw speed. Because people would often ask, well, okay, you're saving me 100 milliseconds here, 200 milliseconds there, or maybe even a second on page load, that kind of thing.

5:19And yet you're claiming also, and people are saying, customers are saying, that they're saving four hours or more every single week. Like, how does that add up? Yeah, how does the math add up? Exactly. How does the math math? So there were a number of early product decisions that seemed very counterintuitive, but ultimately which were spot on. So number one, speed is the number one feature. And everything that you do from loading the page through search to sending an email is blazingly fast. And that's even true if you're on a bad internet connection. Imagine being in the back of a car. That one's usually okay.

5:52Or on a plane. That one is the worst. If you've ever tried to use Gmail or Outlook, but especially Gmail, on a United Wi-Fi, I mean, that's the best one. but like any of them, you're going to be waiting 10, 15, 20 seconds just to send an email. And there's like a 70 % chance it's not going to go through. It was really, really painful. And the way I would explain this to investors in the early days is by way of analogy to Tesla. So I raised the pre-seed and the seed in the Series A from 2014 through to 2016. And just like Uber and Lyft were having their heyday, Tesla was also having their heyday.

6:26So basically every investor I was pitching had a Tesla. And I would ask them to imagine that feeling of hitting pedal to metal and your vehicle going from 0 to 60 in three seconds or less. And then I would ask them to imagine trading that back for a regular gasoline car. And then the emotional reaction was visceral and immediate. These investors were like, hell no, you know, I love my Tesla. You're never taking it away from me. And I said, that's exactly how people will feel once they've experienced truly fast software, the fastest email experience in the world. And then they have to go back to Gmail or Outlook.

7:02They'll be like, what am I doing? It's going to be that same kind of reality shift moment where after using superhuman, they simply can't but live in superhuman. So that's actually really where the speed came from or why we started with speed. But there are so many other things that add up to make the math math. like number two for example the app is designed for keyboard shortcuts that means there are almost no buttons to click it's actually deliberately a bit annoying to use with the mouse and everything about the app is designed to teach you how to use keyboard shortcuts and that alone saves hours per week and days per year now since then of course we built so much more on top of it but in those early days when we were claiming two or three hours a week it was really from that plus speed as well as, and then the final piece was the power of the best Gmail extensions built right on top.

7:52So for example, with the Gmail extension Boomerang, which predated Superhuman, you could set a reminder on an email. I believe they were the first company to let you say, if this person doesn't reply, then remind me to follow up. But if they did reply, don't remind me, which is great because it allows you to offboard the mental overhead of remembering who to follow up with and who not to follow up with. Nothing will fall through the cracks. Now with Superhuman, And as you know, you can do it at the speed of thought. You can just hit a shortcut, enter a few characters about when it should come back, for example, 2D for two days, and then you can immediately stop thinking about it.

8:26You'll never drop the ball again. Whereas in products like Boomerang and now Gmail and Outlook, you can of course do this, but it's gonna take you over a dozen clicks. And as a result, you only end up doing it on some emails, in which case it's no longer a foolproof system. So those were some of the early product decisions that led to saving hours per week. And of course, the product has massively evolved since then. We have a huge investment in AI, as well as multiplayer and collaboration. But just that collection of features alone was enough to build something significantly better and faster than Gmail.

8:57One of the things that strikes me as really interesting about this, you know, for anyone who maybe hasn't built a product before or built a technology, maybe this just sounds like, oh, great, you know, they built a better email. But you're building this on top of standards, SMTP and whatnot. And as someone, you know, I built a podcast company on top of RSS and wave files. And and we learned firsthand that, like, it's really hard to innovate on top of a standard that has ubiquitous adoption. So how did you guys go about doing this? Like, how did you actually create reminders and how did you create actually faster email when you're at the end of the day, like you're still beholden to the protocol and the standard of SMTP?

9:41Well, fortunately for us, and I'm not sure whether this was the case with Anchor, we were able to deliver a lot of innovation without actually innovating on the core protocol itself. And you don't have to do it that way. There's a famous Paul Graham essay, I think from certainly predating Superhuman, maybe from the 2010 or 11, 12 era, one of the early request for startup essays, where he noted, like everyone knew, that email was broken and that it hadn't changed significantly since Gmail innovated with nearly infinite storage and a threading UI. And so his idea was, well, maybe you could put something in the protocol.

10:18You know, most of us technologists know about email headers. It's a good way to stuff information in there. You could put things like priority or tasks or assignee or delegation or all kinds of stuff you could hide in an email. And even if you didn't want to do it in the protocol, you could hide it in the body of the email somewhere. But we didn't do any of that. All we did was focus on the end user experience to say, you know what? We're not going to change SMTP. We're not going to change IMAP. those are going to continue to work the way that they do but what we are going to change is the core experience of driving your email application how can we make that as fast as possible how can we make that as productive as possible and what not that many people might know is that before i was a founder i was actually a game designer and so i went very deep into the study of what is game design and how do you make things fun to play how do you make them joyful and delightful as well as just fast.

11:10And so for every single thing we did, for example, that reminder case where, you know, going from a dozen clicks to set a two-day reminder to you have the thought and before you even know it, you've set a two-day reminder. That's an application of those early lessons of game design that I was absorbing when I was doing previous stuff. Makes a lot of sense. Yeah, so just focus on the user experience. Before we get to AI and like how superhuman is using it, But congrats on the acquisition by Grammarly. Obviously, incredible accomplishment. I would love to hear the story of that. Tell us, like, how did this come together?

11:46How did these two companies join up? Take us through the play-by-play. Well, I met Shashia, the CEO of Grammarly, many, many years ago, back in 2017. And as you know, back then, he was the co-founder and CEO of a company called Coda. And we were at a conference, and it was one of those very special moments where nobody else was around. so we could get into some really deep conversation. And this was the lobby conference, Lobby Enterprise, put on by, at the time, David Hornick from August Capital, now from Lobby Capital. And he and his team do an incredible job of Q Racing some amazing people. And I think there's a lesson here, actually, for founders that are listening or watching, which is I really did not want to go.

12:30This was relatively early in the history of Superhuman, 2017. So just three years after we'd started. And, you know, we were probably 10, 15 people at most. And, you know, I just felt like everything would fall to a standstill if I were to leave the office. And I just needed to keep on being around. Not the best use of your time. Focus on the product. Yeah, yeah. And it might seem weird. It's like go hang out in Hawaii for four days with 200 other people just like you. It's actually a very compelling pitch, but it was still so difficult for me to do this. Nevertheless, my co-founder Vivek Sodera, who's very well known for being a network extraordinaire, that there was no one he couldn't immediately get in touch with, basically pushed me out of the door and registered for me as I was on the way to the airport.

13:18And I'm so thankful he did, because this and so many other amazing things have come out of events like that. So Shashira and I were at this conference, and we were either early or late, I can't remember exactly which, but just the two of us, sitting by the pool, margaritas in hand. And again, no one else was around. And it turns out, and you know this, he's a huge productivity nerd. Coder is a productivity company. But even more than that, back when he was at Microsoft, he worked on Outlook as a software developer. And so he was pre-wired to be super fascinated about Superhuman, what we were doing and why we were doing it.

13:53So I show him the product and it was really early, but he said, oh, I really want this. Can you give me access? And so I onboarded him. and as you know back then we only did one-to-one VIP concierge onboardings so I onboarded him right there right by the pool and as those who've gone through an onboarding would know one of the very last steps is when we ask people to close Gmail and when I asked him to close Gmail he was moving his mouse up to that Gmail tab another tab caught my eye which is an app called Krypton now I asked Shisha what is Krypton and he then proceeded to give me the best product demo I'd seen in years.

14:29It was a document, but also a spreadsheet. It was also a database. It was a collaboration tool as well as a mini app builder. And then as we know, Krypton eventually renamed to become Coda. And then of course, Grammarly acquired Coda late last year. Now, when he wrote his acquisition announcement, he said something along the lines of, as I watched the foundational capabilities of AI change, just about how every tool and surface operates, I started drafting my 2025 planning memo for the team. I titled it the AI Native Productivity Suite. And when I read that, my jaw dropped because at Superhuman, our vision has always to build the AI Native Productivity Suite of choice.

15:10Email is a critical part of that suite and a much bigger problem than most people realize. As I said, there are billions of us in the world and we spend north of a trillion hours every year on email. So Shashira and I caught up and over several conversations, it became very clear that we shared the same vision, which is to build the AI native productivity suite of choice. And it also turns out that email sits at the very heart of where Grammarly is used today. For professionals, email is actually the number one use case for Grammarly. Of the top 10 apps that Grammarly is used in, three of them are email apps.

15:4517 % of the words accepted on Grammarly are accepted in an email experience. And Grammarly helps write more than 50 million emails every single week. So you guys discover that you basically have the same vision. Two different companies, two CEOs, same shared vision. And like all the great acquisitions, you realize, hey, one plus one actually could equal three in this situation. We could put two things together and walk away with something greater than the parts. And if I'm understanding your strategy, is it effectively that this becomes almost like a horizontal work operating or productivity operating system?

16:23You've got email through obviously superhuman. You've got docs through Coda. You've got, I don't even know what you would call Grammarly in this case, like the intelligent sort of like cursor that lives everywhere. And then you have sort of AI powering all of it. Is that like a good way to think about it? Yeah, almost exactly spot on. Imagine it, another way of thinking about it is imagine a platform for apps and agents. We think the AI products of EC Suite of the future is going to be this platform for apps and agents. And so far we spent most of our time talking about the application side. Of course, you have Coder, which is really a collection of applications.

17:01You have email, and then there'll be future applications to look forward to as well there. We also are going to have an ecosystem of agents, some of which we built, but this is truly an ecosystem and we think other people are going to come and bring their own agents into this ecosystem because of the distribution. That's Grammarly agents, which is really a proto-agent, one of the very first agents ever made, has north of 40 million users. And we like to think of it as an AI superhighway down which many things could travel. Today it's really only carrying one agent, which is the agent that helps you improve your spelling, your grammar, and your overall language.

17:40But it really could be anything. And this is one of the most misunderstood things about Grammarly. Grammarly itself is this incredible business, makes a ton of revenue off this one agent. But that distribution is perhaps the most powerful thing because that agent now plugs into north of 500 ,000 applications and websites. So you could imagine that agent actually being a collection of agents And then that agent bringing AI to the edge wherever users are working and also us having a series of first party applications for the teams that want the best, the most powerful and the most AI forward applications that exist.

18:18So you take effectively these sort of vertical workflows and then on top of it, you layer Grammarly, which, like you said, touches 500 ,000 more. And when you put it all together, again, you have almost like this, I'm doing like a horizontal kind of almost like an operating system. In thinking about it that way, where my head goes is, wow, your competition and who you're really going up against now actually are the OSs. Like it seems like the types of companies that maybe might try to find a similar offering here are the Apples through, you know, Mac OS or iOS or even Google through the Android operating system.

18:54Is that kind of how you're thinking about it? I don't think so. Okay. It's only the operating system companies you mentioned, Google and Apple, to the extent that they also have an interest and a vested interest in building AI assistants. And some of them, Google, especially Apple, to a very small degree, have some kind of productivity and collaboration suite going on. But I think it's inevitable that any company in the future, in the very near future, is going to be building AI assistants and agents. I'm sure you're seeing this as an investor. It's almost like as weird as asking the question, well, are you a mobile company?

19:32Or before that, are you an internet company? Which is, of course you are. If you're not building an AI agent or an AI assistant in some form or fashion, then what are you even doing? Like that is going to be core to almost every single software company. Now, I think the unique twist that we're cooking up together between Superhuman and Grammarly and Coda is it's not just the agents. It's the agents and the applications. And I think that is very rare. I think we're also seeing other companies' attempts to do this. OpenAI attempted to buy Windsor for various reasons that didn't go through. But I think most companies who are really deep into this game are realizing you can't just have agents running around doing their thing.

20:13You also need to own first-party applications to be truly incredibly competitive. Yeah. And first party data. Right. That seems to be the other the other big battleground right now. We're seeing all these data walls going up. Obviously, companies like, you know, Salesforce and Slack bringing up the data walls. I think I heard Atlassian's doing this as well. And I guess what's interesting about Grammarly is you get this added benefit of kind of both having your own first party data through the applications, through Coda, through Superhuman, but through the Grammarly sort of agent, like you said, lives in all these different places.

20:48I imagine you kind of get data. You almost have this like unique access to third party data as well. Is that right? Absolutely. Yeah. So you really do need both. I mean, imagine answering the exam question, how would you get the most access to first-party data? And then also, how can you use third-party data wherever it is? And then also, how can you then output that in the most useful fashion? So the applications are what gives the set of agents knowledge about you, your work, the organizational knowledge, and how to pull that all together. So for example, just between Coda and Superhuman, you have the wiki, the documents you may have made to the extent that you're using it for spreadsheets and databases, all of the information there.

21:33Plus then with Superhuman, your email, your contacts, your calendar, and very soon you'll be able to connect other sources of information directly to Superhuman. In addition to, now you add that to the Grammarly agent, wherever you happen to be working, what's on the screen at that time. It's already absorbing everything on the screen, running it through an LLM, giving you recommendations. Now, you put all of that together, and I think you have an extremely compelling AI productivity suite, or as you even said, an AI operating system that helps you be, we believe, like an order of magnitude more effective.

22:07So what are the other verticalized workflows you need to own in this operating system? Again, you've got docs, you've got email. What are the next ones? You mentioned Calendar. Calendar plays a role as well. I guess there's got to be others as well. If you're going to build sort of like this new AI workspace of the future. I think we're still figuring out the order. So I'll give you a little bit more of a zoomed out answer, which is, you know, often people ask me, well, you know, why did you sell Superhuman to Grammol? You didn't have to sell Superhuman. You know, for example, we would have been breakeven this time next year.

22:44We grew twice as fast last year as we did the year before. The real answer is how exciting the combination was. It was like you were saying that one plus one equals three approach. And with this deal, we're just going to have access to significantly greater resources to pursue the vision that we want to pursue and Grammarly also pursuing that vision, which turns out to be the same vision. So we're going to be investing even more than we've done so far, which has been a lot. So even more than that, in our core email experience and in AI, we're going to be building out calendar. We're going to be building out tasks.

23:16And if you're in Superhuman today and keeping a close eye on it, you'll know we're very, very close to a fully featured calendar experience. And we're going to connect all three of those things beautifully together. I will then reimagine chat and redefine collaboration. And that will pull on everything that we've learned about work communication over the last 10 years. And then, like we've been talking about, we're also going to create this new way of working with AI agents. And we think the big upside there is that these agents are going to free all of us up to be more creative, more strategic and closer to achieving our human potential.

23:50So hopefully that gives you a little bit of a glimpse into what's to come there. So one of the things that keep bouncing around inside of my head as I hear you tell this story is something that Shashir, the Grammarly CEO, Coda founder, has spoken a lot about to me and I know to others on the Internet. And that is the concept of bundling. And Shashir's got this, you know, somewhat famous essay about sort of the benefits of bundling and the strategy for bundling and, you know, super fan overlap and weak fan overlap. I don't know if you know what I'm talking about. We'll drop a link in the notes here.

24:25But how much of bundling and bundling theory factored into this? Because, again, it seems like now you're taking these verticalized workflows and bundling them together. And yeah, it strikes me that some of these have very, very strong overlap in terms of their potential users. And some of them have have weaker overlap, which actually turns out to be a good thing for bundling. So how much of that is the strategy here? Yeah, I think it was a pretty big part of the strategy is to create this really, really compelling bundle. And then for those that don't know, Shashir is the bundling master. The essay, I believe, is called The Four Myths of Bundling, if I'm getting that right.

25:00That's right. The Four Myths of Bundling. Yeah. You can read it on Coda. It's a really great Coda essay. Whilst you're there, you can read my Product Market Fit essay, another really great Coda essay. And I'm going to do my best. I'm not the expert. She's the expert. But let's see if I can help explain some of the thinking behind this. So you can take any of a product and you have strong fans and casual fans. And a strong fan is someone who would buy it a la carte independently at like the individual retail list price. and a casual fan is someone who is going to appreciate it if it's part of a bundle.