Ep 6: Jasper CEO Dave Rogenmoser on the Future of Writing with AI

Jacob and Erica sit down with Jasper CEO Dave Rogenmoser to discuss the future of writing with AI and what it means for marketers and the internet, Jasper post ChatGPT, hosting a massive Gen AI conference, going upmarket and more. You can find Dave on Twitter (@DaveRogenmoser) and learn more at Jasper.ai

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:05Jacob Effron?

  This is Unsupervised Learning, a podcast on the rapidly developing AI landscape and what it means for businesses in the world. I'm Jacob Effron and I'm joined by my colleague, Erica Brescia, and we have a really exciting episode today. We are joined by Dave Rogenmoser, the CEO and co-founder of Jasper. Jasper is a gen AI platform for businesses to create content, including social media, advertising, articles, and more. The company recently raised $125 million at a $1.5 billion valuation with investors like Insight, Bessemer, IBP. So a real success story in actually getting these fascinating models that are out there into the hands of users, really starting to derive value from those products. So, Dave, thanks so much for coming on the podcast. Really excited for the conversation.

• 0:46Speaker 01

  Yeah, we're getting ready to rock and roll here.

• 0:49Jacob Effron?

  Amazing. Well, I'm sure most of our listeners are familiar with Jasper, but for those who are not, I thought at the top would be helpful if you could just give a brief kind of context on the product, how it works, and how most people are using it today.

• 1:00Speaker 01

  Yeah, I mean, your intro was perfect. That's almost word for word how I described Jasper, but we're basically a platform-

• 1:06Jacob Effron?

  Almost like I took it from the website.

• 1:08Speaker 01

  Yeah, exactly. We're a platform that helps marketers and marketing teams write high-quality content faster. We started about two years ago, two years and some change ago, really before kind of all of the madness that's happening now, quite a ways before that, but just saw some cool technology, saw that AI was at a place where it could actually write pretty good content. And just as a marketer by background, a copywriter by background, I felt like if I was excited about it, and if I thought I could use it for myself, even though I felt competent to actually go write, then like there's this massive long tail of people that probably feel less excited about writing content than me, and that if we could build a simple interface around it and then build great

• 1:52Speaker 01

  community and great go-to-market, then we could get this thing out there to the hands of regular people. And that's basically what we've been building over the last two years.

• 2:00Speaker 02

  And while Jasper's been the focus for two years, I think I always loved the founder journey. And you've called it like an eight year overnight success story. I think you went through Y Combinator in 2018. I went through in 2013, by the way, and I have a similar like overnight success story. But we're here to learn about yours. Can you talk about the journey leading up to Jasper?

• 2:19Speaker 01

  Yeah, so I started building companies, companies, I'll put that in quotes because these were not companies. But I started trying to figure out how to not have a boss with two of my friends back about nine years ago. They're still the co-founders of Jasper. We've been sticking it together, you know, through four or five different iterations of company. And I think if I look back in hindsight, I realized that like each one of those companies played a big role in like what Jasper is today. And we, we learned a skill set or developed something about ourselves or our team that gave us an unfair advantage in Jasper and allowed us to like hit the gap hard, you know, when we saw the opportunity here. And like, I just absolutely would not have been able to do Jasper without doing, you know, really all of those.

• 3:01Speaker 01

  So it was like, it was a marketing agency, turned into like a course, coaching business, turned into Proof, which is the company that we, we did Y Combinator with, which is our first, you know, real like software startup. And each one of them was like a little better than the last. Each one of them, we learned a little bit more. Each one of them, we just figured out how to, you know, one more step of scale. But none of them were like really great. And so then Jasper comes along and everyone's like, you know, who's this team? Where'd they come from? You know, it's just, you know, oh, flash in the pan, you know, AI is just gonna make everyone a millionaire. And, you know, and for me, I'm like, it feels like one big, long, connected journey. And, you know, all those things add up.

• 3:38Speaker 02

  Yeah, you're kind of like in training, right? Just for this moment. And then once the moment is here, it's like an athlete or something, then you go, go, go, go, go. That's awesome.

• 3:47Speaker 01

  Yeah, it really feels like that. I just felt like, oh, like we know how to build software. We know how to do community. We know how to run paid ads. We know how to do copyright and we know how to do like all, you know, we've worked together in like high stress environments, you know? So like our, our co-founding relationship is like strong. And it's just like all these things that, you know, any one of those is like a chink in the armor. We had just built them through really Really hard times when things weren't working and then Jasper's kind of the first thing we did that that really worked well.

• 4:15Jacob Effron?

  Yeah, I mean, speaking of being ready for that moment, I feel like there is this like, it feels like this monumental moment where I believe you like sent your co-founder JP a text about like GPT three, you'd been exposed, you were like, this seemed pretty exciting. Maybe just talk us through like that moment, you know, what kind of made you realize you were on to something there and inspired that?

• 4:31Speaker 01

  Yeah, I mean, we've always, you know, like I just shared pivoted a lot and I've always believed in the pivot. And that if you can just kind of take a piece of what you're already doing, but then kind of go find another piece that you're not doing where it's so it's not like totally unfamiliar. I'm not getting into like starting hospitals or something. It's like, you know, it's okay. It's like marketing and I know that, but then it's this like new AI thing that I don't know. You're in pretty good shape and we had just kind of decided to spin down proof. We weren't going to work on it anymore. We'd laid off, you know, a lot of the team. And we were just kind of figuring like what was next and not in a big rush. We knew we wanted to be really excited about it. We knew we wanted it to be, you know, more powerful than what we were doing with proof. And then, yeah, I had just seen, you know, some tweets about GPT three and thought, oh, this is really cool.

• 5:15Speaker 01

  And I tried to get access to it, you know, the summer, uh, summer of 2020, maybe when it came out, couldn't get access to it. And then it was like October, I found out like why combination companies could get access to it and just like, you know. And it was one of those things where I just had this gut feeling where I was like, this is gonna work. And now I had no idea it would work as well as it has, but I was like, oh, this is gonna work and we could probably get to like, you know, a hundred thousand dollars MRR and, you know, whatever, like, you know, a small goal. But I was just like, it felt like everything I had learned and knowing the customer, knowing software, like all of it made sense in my mind. If we could just take this out quickly before like the world gets, you know, it gets commoditized and super competitive, then we'll be in really good shape. And so, yes, it's like one of those late night texts to him. Luckily I like found that text and screenshotted it and tweeted it out.

• 6:00Speaker 01

  It was just like, dude, this is it. It's time to go.

• 6:03Speaker 02

  How surprised were you when you started playing around with it, it just how good it was out of the gate?

• 6:09Speaker 01

  Yeah, I mean, the first day I played with it, it was not good and I didn't know how to use it. I didn't know anything about prompting. And so I remember like the first like four or five hours, I was a little bit like, all right, false alarm here, you know, like this is like, you know, those might've been demos we had seen or whatever. This is like, not that good. But then like, we really figured out how to talk to it, how to prompt it, how to, you know, and again, a lot of it was, you know, giving great marketing examples that we had, again, I had a whole swipe file on my computer of like, really great marketing copy and copywriting, you know, from past campaigns that we'd run and all of that. So we kind of figured out how to do it and it was pretty magical. And I even remember those first like Zoom calls with early prospects. This is before they, you know, they couldn't use it. I could barely use it, but I just would share my screen and be like, hey, like here's our website headline generator.

• 6:54Speaker 01

  You know, tell me about your company and I'll type in a little box and it would like spit out a headline and they would all just start cussing and being excited and being like, when can I use this? Cause like, I will drive to your house right now. If I can like get on your computer, start using this for my business. And that was just a very different feeling than anything else we had launched. It was always a little bit more like, I'm kind of trying to get people to interested and I'm kind of pulling friend favors and you know, it's like, this was like, okay, people are like beaten out the door for this thing. So even very early days, it was, it was incredibly exciting.

• 7:23Jacob Effron?

  Yeah. And I feel like you and you and your customers are kind of some of the highest users of these models today for actual business applications. And so I'm kind of curious, you know, even with today's models, like where are you guys kind of noticing that they work well, maybe where they fall short? And I think our listeners would be super interested in even just how that's changed over time. Obviously the underlying models have gotten a lot better, you know, since you guys started the company would be super curious for your perspective on that.

• 7:45Speaker 01

  Yeah. Early days, I described Jasper as being really good at use cases where you had to write a lot of words and they had to be pretty good, but not great. And so with maybe a, you know, whatever, an email to your like marketing list, you know, it's kind of be like pretty good. And so like, you know, it might struggle a little bit more there, like very early days and that's not true anymore. But early days, you know, but like, let's say like a blog post for your website, you know, and your team's going to crank out, you know, 30 blog posts this month and like, they've got to be pretty good. They got to hit the mark, but it's Like it's okay if you have like a weird sentence or two, like inside of a blog post or something like it's like a little fluffy or whatever like, I think it's not ideal, but it's, it's okay.

• 8:26Speaker 01

  And you just want a lot of words like that's the kind of stuff that Jasper, you know, was really good at, you know, early days or even use cases where you just needed a lot of variants. And so let's say like Facebook ad headlines is the kind of thing where you just, you know, nowadays you just load it up into Facebook and you'll just stick 10 headlines in there. Let the algorithm figure out which one's the best. And then you need 10 more the next day and then 10 more the next day and 10 more next day. And it's like, you know, everybody ultimately kind of gets jazzed about it. You know, you're running your ads, you do that like one or two days and then you just like, let it run. You're like, I don't have time to be thinking of like new things all the time here. I like Jasper is really great where you could like pop in your winning headline and say, hey, give me like 10 different variants of this. And different tones of voice and different angles and different benefit statements and all of that. And it's like really good, you know, it could do that infinitely.

• 9:09Speaker 01

  So it's like another place that like Jasper worked really well out of the gate. And yeah, we've just seen the models improve dramatically over time. I mean, they can write much longer in one shot where back in the day it was just writing a sentence or two. Was kind of all the most could do before they would start to really like tail off and wouldn't be as useful, you know, even writing a long paragraph, you know, the last sentence wouldn't be, you know, as good as like the first couple of things. You know, we see them, a, a big problem, and this is still a problem, but we see them just, you know, say incorrect things and lie about things and do it very convincingly. That, that was a really big problem. It seems like it's gotten less and less and they're actually getting better, like staying, you know, on track and not hallucinating about things.

• 9:50Speaker 01

  And another thing that I think a challenge for Jasper, and you've seen us fix some is just, just fact checking. And like, you know, Jasper would work really well for use cases where it was like very commonly known. It's like, let's say like writing a blog post about like growing Bermuda grass. It's probably not going to like hallucinate too much in that kind of thing. And it's low stakes, right? Like if you like, say that it should be cut at four inches instead of three inches, like whatever, not the end of the world.

• 10:15Jacob Effron?

  Depends how much your end customer cares about their lawn, but yeah.

• 10:17Speaker 01

  Yes, totally, totally. But like, you know, writing a blog post about, you know, the life of Martin Luther King, you know, is like probably not going to go that well. And you're going to have to like really fact check that. So we, you know, we've always told Our users to fact check everything. We've always told our users that like the best content is going to have a thoughtful writer. Very engaged in the whole process. This is not just something where you go click a few buttons and get something great. I think that's still true. That'll likely always remain true because like as the bar gets raised, just, you know, everyone can write a great blog post with a click of a few buttons, but if you are really thoughtful about your audience, you really think deeply about the message you want to communicate and you can do it in a creative way. You're going to stand out. You should just use these tools to do that.

• 10:56Speaker 02

  How does the content of the internet change when you project forward? Given how easy it's getting to generate just massive amounts of content, even if you have a human editor, if it takes you 10% of the time or something in the future to write, like, how does that change the internet?

• 11:13Speaker 01

  I think probably dramatically and you know, there's kind of a joke around, we're going to have this, you know, basically AI protocol where, you know, I'm going to take a sentence and expand it into a full email. Send it to you. You're going to have AI on the other end. Take my full email and, you know, summarize it into a sentence and like, we'll never actually know what all the AI is talking about, you know, and the in-between. And so I don't know. I think there's going to be a lot of that where there's a lot of, you know, generating long content, great content, hopefully using these tools to write better, better content. And like, you know, obviously we've got billions of blog posts out there right now. But, you know, you all have felt the problem when it's like, it's kind of a niche question. And it's just like, hey, like, what should I do this Wednesday in this city for this budget?

• 11:57Speaker 01

  Like, there's probably not a blog post that like really answers that. Like, there might be, you know, blog posts like that in the future. Or just even, I think you're going to get, you know, obviously you see this like chat GPT and you see this like, you know, and being embedded more in the search models too. You're going to see just much more summarized versions of all of this content out there. And that's really going to change the game. There's even a world where like, you know, maybe blog posts are not things that people go and read themselves as much, but it's, it's almost acting as just training data that a AI assistant would go and crawl for you and go pull from. And you could still get that information, but you would get it in a different format than going and reading a 5,000 word blog post there. But no, it's certainly going to change content a ton and I think it'll be for the better, though it's going to be very different.

• 12:40Jacob Effron?

  Yeah, it's super interesting. And you kind of talk about this process now with these tools of you're kind of recommending folks, you know, still, you know, not just auto generate the posts, but be involved in the writing process. I wonder, like, do you guys at Jasper have like, uh, you know, I'm sure as people learn to use these tools, they're wondering what the right new workflow is. Do you guys have like a suggested way of like, hey, maybe like generate a draft and then, you know, come in here or like, how do you think about what like a writer's workflow in the, in the future looks like?

• 13:04Speaker 01

  Yeah, this has been a big challenge for us over the years because, you know, this is a brand new way of working for everyone. And, and we've got to kind of balance giving them familiar workflows. So, you know, our, our document editor like looks kind of like a Google Doc. And, you know, it's, it's kind of got the same feel and you're going to feel at home being in there. But then we've got a compose button and we've got the ability to highlight and rewrite. We've got some other things that are brand new and, you know, we've, we've interviewed so many users trying to figure out like, hey, like, what is your ideal workflow? And it turns out like, I mean, obviously everyone has a very different way of working. And so, you know, we've tried to balance being sometimes very opinionated about it and that the workflow is a little more rigid.

• 13:47Speaker 01

  And we'll build these workflows, you know, we've got a workflow tool where, you know, it's kind of step by step, how to build a blog post or how to write an email or how to write social media posts or whatever. And it's not very flexible, but it's very, it'll guide you down the whole process there. And then we've also got the ways that you could just, you know, through chat or in our document editor, like very free form do it. So I don't know, we're always going back and forth on, you know, how do we make this flexible? But that's confusing for like the brand new user or very guided, but, but not flexible. And I think for us, a lot of it is giving them more opinionated workflows and guides and templates to start while they kind of get the hang and like learn about this. But then like let the product unfold over time to where, you know, as they use it more and more and they get the hang of it, that they're really using it in a very unstructured way.

• 14:31Speaker 01

  But yeah, I don't know that we've ever nailed the balance there, but it's certainly something we think about a lot.

• 14:36Jacob Effron?

  Yeah, it makes a lot of sense. And then one thing I'm curious about over the course of the business, I mean, obviously when you started, like you said, it was hard to access GPT three, you know, you, you guys had got access to Y C company. A lot of folks didn't have access. And then I imagine it was, it was a pretty wild time when chat GPT was kind of introduced. And those first few weeks, I'm sure were crazy as, as, as other folks were playing around with it, like, would love to hear kind of what those weeks were like, you know, how that's kind of impacted Jasper and, and, and, you know, your thoughts on that.

• 15:01Speaker 01

  Yeah. So I mean, for the first year and a half of the company, you know, no one was really thinking about AI all that much. Even when we raised our funding round, you know, we were, we were early days and there wasn't this like AI boom. And then, you know, I think stability really stable diffusion, you know, sometime like September of, of last year. And that seemed to like really kickstart a whole bunch of stuff, a whole bunch of people working and building really cool stuff. And then chat GPT launched, you know, end of November, you know, of last year, obviously. And, you know, I didn't, I didn't think a ton about it at first, you know, I saw it and thought it was like a chat bot model. And I was like, eh, we don't do chat bots. And so, you know, my, my very first reaction was like, all right, I got other stuff to do today. Like I'll, I'll go look at this thing tomorrow. And then, yeah, obviously like it caught on and I think what they nailed was that chat is an interface.

• 15:46Speaker 01

  That is, it's kind of like what I just said, like it's a familiar enough interface that people kind of get what to do. But it's also like flexible in that it allows you to like go very broad. And obviously it's like a very like general tool and, you know, can write poetry and can write code and can do like, you know, all these things there. But it's like, I think that chat interface was something that, that people understood really well. You know, for us, you know, we saw like the market come alive in a really cool way where, you know, larger companies didn't want to talk to us, you know, a year before that. And, you know, you would have gotten fired if they found out that your marketing team was using AI to write content. And then just in about 30 days, it was like, you're going to get fired if you don't use this. And so all these companies, you know, that were previously interested were now at the table.

• 16:30Speaker 01

  And they were saying, hey, we need something that is great for teams, that's secure, that is built for businesses, that helps us do our work faster. And Jasper looks like, you know, they do that. And so, you know, I think a lot of people, you know, try chat GPT and it's obviously a terrific tool. But, you know, again, it's, you can't collaborate on there. And there's a bunch of kinds of things that more like larger businesses, you know, would, would want to have their teams do. And so I think for us, it really brought them to the market, shifted our focus up market more. Because we saw that the bottom, you know, get more commoditized and, you know, this was like this good free tool that I don't want to compete against good free tools. If I can help it. And so, you know, I think we just got more focused on teams, more focused on businesses, more focused on the up market. And like thankfully, we'd like Been in the space for a long time, had a great team, had, you know, great cash at the

• 17:14Speaker 01

  bank and was like, you know, easy to kind of pivot, you know, into that. And so it's been, again, I think harder at the bottom end and like will continue to be, but like much more blue ocean and availability at the upper part of the market.

• 17:26Speaker 02

  And, you know, as investors, we're always looking at the competitive dynamics of a market, right? And what's how it's going to evolve over time. And you make a really good point on the like, you know, the low end of things. It's gotten a lot easier. There's a lot of great free tools, you know, as these models continue to improve and more companies are getting funded in this space. Like, how do you think about the long-term opportunity for differentiation for Jasper?

• 17:50Speaker 01

  Yeah, I mean, I think there's in every business, a lot of places to differentiate, you know, just from the, let's say from this AI perspective, you know, we have a lot of great data from our users, how our users are using it. And I really think fine tuning is a real opportunity for companies. Everyone should be thinking about how to do that and how to take these general models, make them specific to your users. The more good data you have, the more sophisticated you are and fine tuning around that and like, you know, making the, you know, kind of the, the cost benefit, you know, analysis and those trade-offs there. You're going to get improved models and your users are going to be happier and you're going to be able to serve them better. And so, you know, we're fortunate enough that we've got a lot of great data and we've got a great team that can go do all of that. You know, that could even lead us to building our own models, you know, someday. I'm open

• 18:35Speaker 01

  there. You know, I don't want to go do stuff that, you know, others can do, you know, better and faster than us. But if we get to a place where it's like, you know, like people just aren't serving our customers in the way that we're trying to serve them here. And, you know, there's opportunity for us to go build some of our own models, then we'll do that. You know, I just think from the product perspective, kind of outside the AI, like knowing an end user and owning that customer relationship is gold. And this is the thing that certainly people missed early days, and I think are still missing now, though less so, is what does this actually mean for everyday people? And how is this actually going to make people's lives better? And that most people still don't know what a parameter is, or, you know, most people still don't know what ChatGPT is. As hard as that is for our world to believe, like the most of the world doesn't know what that is. And like, they just have

• 19:20Speaker 01

  a job to Do and a use case to be solved. And if you can help them do that, then they're going to love you and be really excited to work with your company if you can do it in a way that makes sense financially for them. And so, you know, I think like that last mile of AI is really hard, really intricate. There's tons of different workflows and ways to do it and think about it. But like staying really close to the customer gives you a huge advantage and like, and like those conversations with customers create IP for you. Like what they tell you, if I went and interviewed a hundred of our customers, like I would have a unique insight on the market that nobody else has. And from my perspective, most people are unwilling to go and really do that, even though it's like easy and fun and free, you know, to go do that there.

• 20:00Speaker 01

  And so, and then on top of that, you know, there's, there's community, there's distribution, there's building just a great organization, you know, those are all like hard things to do. But I think like there's plenty, again, if you get focused on the end user, try to solve a specific problem, don't go too broad. You know, even just like the last six months, we've kind of narrowed our focus, then like there's, there's so many great businesses to be built here for the people that'll go do it.

• 20:21Jacob Effron?

  Yeah, it's super interesting. You kind of talk about this like dual impact of, of the explosion of interest in AI where maybe, you know, maybe the, the bottom becomes a bit commoditized. But at the same time, like everyone has to use this now. And so it's like, I'm sure, you know, you're, you're able to get in the door of pretty much any, any company you want to talk to. And so it sounds like overall, maybe, you know, Jasper as a company has shifted a little bit more towards, as you were saying, like the teams and businesses and upmarket. I'm curious from like a product perspective or, you know, what does that mean, you know, in terms of kind of what you guys are focusing on or, you know, some of the different requirements, I guess, for some of the larger enterprise?

• 20:51Speaker 01

  Yeah. And, you know, this is all just comes out of customer conversations, you know? So like, we always try to be pulled out. So this is not, you know, me being some, you know, guru that can see into the future. It's like you go talk to like this new set of customers and you're like, hey, like, where do you like the product? And they love, they like these X, Y, Z, and they're like, what's fallen short? And then they tell you and then you can go build these things. And so, you know, for us, it's like, you know, Jasper used to be more like solo player, single user. And like now we're, we're allowing people to just add, you know, more of their teams on there and collaborate, you know, across workflows on there. Jasper has been, you know, unpersonalized up until recently where it doesn't know anything about your company, it doesn't know anything about your brand voice. And now we're allowing these companies to upload.

• 21:31Speaker 01

  Style guides into Jasper and Jasper will follow those upload company information that Jasper can go and reference upload product information and just like Jasper can learn about those things. You can also set like a brand voice across your whole workspace or set the brand voice that you want to have. You know, I think we're finding is like a lot of people like think they have a brand voice and you're like, this is kind of the same as like everything else, but we could like make it 10% funnier and we could like, we could give you a brand voice for the first time. And then make sure that everyone writes in that like that is really exciting. And then finally, just, you know, Jasper, like we want Jasper to be everywhere that you're working. And so, you know, any marketer at a company is using, you know, 10 different tools. You're writing, you know, a blog post and you're going into Facebook ads and you're running the writing an ad to promote that.

• 22:15Speaker 01

  And then you're going into HubSpot and you're sending in an email and like, we want Jasper to, to be in all of those places in a very natural way. In a way that knows what you just wrote in a way that knows about how you wrote it in a way that knows kind of what style you want to write that in. And so, you know, we've built this great, you know, browser extension that, you know, takes it a lot of places there. We're rolling out an API that can help do that. And so we're just trying to get Jasper like out of the web app and into the places that you're already writing. I really think that's probably the default. I think that'll be the default state, you know, in the future and it'll be less kind of go to jasper.ai and do your work there. I think that'll happen some, but I think it'll more just be like, hey, Jasper's like where you already are.

• 22:52Speaker 02

  The brand voice thing is really interesting though. And I think, you know, in this new age of almost fully distributed and remote work, it actually makes it a lot harder to kind of onboard and ramp up marketing teams and get everybody speaking in the same voice. You know, just AI aside, the world has changed a little bit in that regard. And now you have this really powerful tool to layer on top of it. For those who haven't seen it, like, what does the process of tuning your brand voice in Jasper look like? And who's controlling that? Like, how does it show up in your work?

• 23:25Speaker 01

  Yeah, I think you're right. I mean, if companies are really sophisticated, they'll have a style guide, you know, that they'll put the effort to put in and then hand it out to all of them. And they're like, hey, please follow this. And if you're really, really sophisticated, you'll actually hold people accountable to it a little bit. But it's like, yeah, who has time to like really do all of that? And mostly it's just kind of an exercise and futility. But yeah, I mean, we just built a different module inside of Jasper, you know, called Jasper Brand Voice. And you can go in there and again, you can, there's, there's kind of several ways to do it. You could, you know, work with Jasper to like, figure out the right tone of voice and save that and like, save that, you know, across the whole, you know, or you could upload documents about the company and say, hey Jasper, like we want to reference this, you know, if I say, write something about my company, now you know about my company because, you know, maybe it wasn't out there. Maybe it wasn't in like the base training models. And so Jasper can reference that.

• 24:10Speaker 01

  So it's kind of a mix of, you know, saving, you know, again, kind of brand voice, but then like these memories or like historical data that you want Jasper to reference there. I mean, it's just all model, you know, like a different module inside of Jasper. Usually admins will set or, you know, some, some level, high level of user will, will set that. And then that'll kind of roll out across all the different members inside the team. And even if you're like a, let's say a marketing agency, you know, you could have, you know, a different brand voice for each client. And whenever you're like writing, you can just go select, hey, which client is this for and go write in that. Or let's say each, you know, different products, you know, have a different kind of brand voice or different contexts you want to include, you can just select that as well. So it's really cool. It's really exciting because this is really has not been able to be done up until really recently.

• 24:49Speaker 02

  Yeah, no, it's huge. I'm reflecting on my time as a founder and we'd hire somebody new. And I just remember how many revs we would go through to try to get the right output as you kind of train and ramp up those folks and the voice that you want to put out there. It's incredibly powerful to be able to do that in a more scalable and programmatic way. Super cool.

• 25:08Jacob Effron?

  How often have you adjusted it for Jasper itself with your own marketing campaigns?

• 25:14Speaker 01

  Yeah, I would have to ask our marketing team, you know, again, I think we've, we've always tried to be playful and fun. And, you know, just honestly, just have fun and like not be, I've probably said a thousand times, like, we're not just going to be a regular B to B company, you know, like we're going to like feel more consumery. But it's even interesting, like, you know, even like in this. Shift up market, you know, we've had to like really think like, okay, like, how do we be fun and playful, but not childish? And, you know, is there a way to kind of move up market and keep a lot of that, but also maybe put a little bit more of a, you know, put our suit code on, you know, as we kind of move up market. And so like, there has been conversations in like Jasper's new brand voice. And we've used Jasper, you know, one, kind of help identify that, but two, you know, helped actually like roll that out across the team.

• 25:59Speaker 01

  And so we're kind of experiencing it and dog food in it like firsthand as we like just make a company shift.

• 26:04Jacob Effron?

  Totally. I mean, one thing I'm interested in is I feel like you guys were one of the first, you know, really successful Gen AI applications. And so I think, you know, people were pointing to you early as like, you know, hey, here's proof that this can work. And I think you guys recently threw this like Gen AI conference, right? You know, convening folks together and, you know, I'd love to hear like what inspired that? And how you think about generally like Jasper's role in convening a community of this next generation of folks that are trying to build stuff in the space as as successful as you have.

• 26:28Speaker 01

  Yeah, I mean, community has always been a big part of Jasper, you know, and it's traditionally been, you know, early days, it's kind of just a Facebook group that I would just hang out in and talk to customers in. And, you know, it was, it was really powerful and still is. So I think we like community, you know, in general. I think I saw this gap in the market where all the AI conversation was very intellectual and academic. And, you know, there might be conferences on, you know, for AI researchers. And if you were just a business leader trying to figure out how to bring this into your company or how to build a product around this. Or like, what does this mean for me as a non-AI person? Like, there really wasn't anywhere to go and there wasn't any conferences.

• 27:08Speaker 01

  And we had one of our investors from Co2 come to dinner, I think it was like mid-November of last year. It might've been like end of, end of October, but you know, right around then. And he was just like, you guys got to host a conference, like the moment is now. I was like, yeah, yeah, we're gonna do one next October. We'd always got a plan on doing one at the end of October of 2023. And he's like, no, no, no, no, it's like now, or it's like, you know, it's not October, it's just now or never. And I was like, oh, well, that's a good point, you know, let me go talk to the team. So I like went and slacked the team on the drive home. And it's just like, hey, like, I think we can do this thing in February and, you know. After some hemming and hawing, you know, Megan, our VP of Marketing was just like, all right, let's just do it. And so she put together this whole thing in like two and a half, three months. And it ended up being terrific. And so, yeah, I just think we just wanted to kind of plant the

• 27:51Speaker 01

  flag and say, hey, like, we were, we were first here and like, we're willing to also kind of be a place for everyone to come and figure out what does this mean for their business. And then really the theme for that was less generative AI, but more like generative AI applied to business use cases. Like, we want to lead that conversation.

• 28:08Jacob Effron?

  Yeah. And you guys sold out. So and clearly there was, there's a lot of, a lot of demand for that. Is it, is it going to be a repeat performance or?

• 28:13Speaker 01

  Yeah, we want to do it every year. You know, I mean, obviously the space moves so fast. Like, you know, it's hard to even say exactly like what the exact, you know, theme of it would be, you know, a year from now. But yeah, we, we ended up selling like 1,200 tickets. And, you know, like, it's like, one of those things, like, a week after we launched it, you know, we had like 40 ticket sales. And I was like, oh no, like, this thing is going to be empty. Like, how small can we make the room where it still feels normal? Like, we had like, bring in the like, the like, the curtain walls to like, I was like, this is gonna be like a fireside chat. And then like, the last like, two weeks, I mean, we like, sold out. I had, you know, CEOs of Fortune 500 companies, like, texting people, be like, hey, can I get a ticket? And I was like, holy cow, like, what is going on? And it was like, packed and everyone just had a really great time. So I

• 28:56Speaker 01

  don't know, it just goes back, like, it wasn't that we did anything super amazing. It's just like, if you are, are able to get in early on a big macro shift that's happening and like, executing that, like, it's way easier than, than doing anything else.

• 29:09Speaker 02

  Yeah, timing is every, I mean, there was that conference here, like, a week or so ago, right? And it was pulled together in like, two weeks. And there were just thousands of people and everybody begging for

• 29:17Speaker 01

  Oh, yeah.

• 29:18Speaker 02

  Tickets for that, too. I think, um, any AI event you put on, you can get folks to. But hopefully you're expanding capacity already for next year, because I'm sure you'll have several times the interest in the event. And to your point, things are moving so quickly. It's You could probably do it every six months and there would be enough net new content and changes that people would still come out and want to learn. So, you know, on this rate of change, you know, I read you're a father, you've got several little boys. I do too, and I'm excited about AI. But at the same time, I find myself thinking a lot about what his job might look like. My son's 10. You know, what are the jobs 10 or 15 years out? And what

• 30:01Speaker 02

  does the world look like for our kids? How do you think about that for your own children?

• 30:06Speaker 01

  Yeah, it's certainly exciting and a bit scary. I mean, mine are mine are five, three and one. So mostly I'm trying to keep them alive till tomorrow. You're like, you're like, you're doing like college prep right now, you know?

• 30:20Speaker 02

  Yeah, not quite yet at 10, thankfully, but he is learning Python and like getting into coding. And I'm like, boy, this is not gonna be like this.

• 30:27Speaker 01

  Totally, totally. You know, I go back to just the basics, you know, with our boys and just thinking about that. It's like, how to how to problem solve, how to think critically, how to communicate clearly. And like the mediums, you know, will change. But like even we have seen, like, I think there's this thought that, okay, AI compresses creativity in people down to where like, everyone just kind of writes the same thing now and everyone. But like in our community, at least, like, we've still seen like a very huge spectrum of like creative thinking and good ideas. And like, our best Jasper users are so much better than like our kind of neutral or worse Jasper users and that they just they use all this technology to create

• 31:09Speaker 01

  really remarkable things that I never could have before. But it's like, the spectrum is still there. And so I think, you know, for for people thinking about the future, it's like, you know, one, like, learn how to use these tools. Try not to do things that are definitely going to be obsolete, you know, in a year, two years, three years. Like, I wouldn't recommend trying to be only a blog post writer, you know, for a long period of time. Like, you might have like a few more months of that. But it's like, if you can kind of step up a level or two levels and think about how to how to just do more than that. And again, like, really understand people and how to communicate with them, then like, Like these tools are just gonna to give you superpowers here. So I don't know. I mean, the rate of change has been much faster in the last four or five months than even I expected.

• 31:54Speaker 01

  I mean, it's, it's so much faster than the first, you know, year and a half of Jasper. Even that was like pretty fast. But yeah, I don't fully know kind of where this is, where this is headed here, but I, I'm an optimist. I think this would be a good thing, but I think people should be paying attention and using these tools. And trying to like, take the next step on top of them.

• 32:10Speaker 02

  And when you, you know, you made this comment about the folks that are best at using Jasper and like the kinds of remarkable content they can, they can come out with. What sense do you have for what kind of productivity boost they get out of Jasper that, that you know, your top few percent of users?

• 32:27Speaker 01

  Yeah, we've got, we've got some stats on this. I'm forgetting the exact top of my head, but yeah, generally, you know, teams are probably 20 to 50% more productive is what they'll report. You know, blog posts that used to take them eight hours, you know, will take them six or four hours, you know, you know, to write a really, that's like a really high quality, you know, piece of content that they're putting out there. You know, marketing campaign that used to take two people to put together for a week, you know, might just take one now. And so it's dramatic. It's not, we're not talking like five, 10% here, you know, it's cool, but it's like, it's, it's, it's dramatic.

• 33:01Speaker 01

  And I still think we're so early days and people even understanding how to use these types of tools and being really good at them. And so, you know, I don't think that means that, oh, everyone is going to be, you know, sitting on the streets, you know, begging for change now. It's like, but people have to adapt and they have to grow. And obviously like marketers are going to be so in demand, you know, for a long, long time. Each one will be expected to do far more kind of per person and hopefully have big outsized results for the companies to take advantage of it.

• 33:28Speaker 02

  Yeah. And you know, when I've talked to some developers who are using, whether it's Copilot or another tool. I've heard from some that they're going from what would take months to sometimes even just a week or two. So they're seeing 80% plus improvements. And what you're talking about from a Jasper perspective is already so material, right? Like 40, 50% is huge. Do we get to 80, 90% like we have with code or is the fact that code is, I guess, more perhaps concrete and sometimes a little less creative in some areas? It's a different dynamic.

• 34:04Speaker 01

  No, I think you will and go far beyond that. I mean, I think you're probably less than a year away from like you type in like a few sentences on some product, you know, release that's coming up and you click a button and then 10 different entire marketing campaigns are built for you in five seconds. And you could have them for every demographic, you know, that you're kind of targeting. So here's the one that reaches boomers. Here's one that reaches Gen Z. Here's your whole swipe file of, you know, 40 blog posts, your, your PR announcement, your emails for your marketing list. Here's 10 Facebook ads. And it could be like for all of them. It's like, you know, that's the kind of thing that would have taken one person eight months to kind of put together, you know, and it could be in like five seconds. I think like we're going

• 34:49Speaker 01

  to go like far, far beyond all of that. But then the issue is if everyone can do that, then like the, the noise level, you know, just like ramps up. It's like, you still have to like, you still have to get beyond that in order to compete. And you still have to have a really good version of that. And you have to really know your audience well. You have to know how to use these tools well. And you have to know how to like align them across the marketing channels. And so it's not just a freebie to have that, but like so much of that work that's a little bit tedious and boring will become easier. And in fact, there's probably people that would be tremendous at marketing, but can't do it very well because maybe they're bad writers. Yeah. So we've got this customer, Jim. He, you know, was a copywriter and an academic.

• 35:31Speaker 01

  About five years ago, he got attacked outside of this grocery store, got hit in the head, really traumatic brain injury, turned into aphasia to where he had a hard time recalling words and communicating. And so he was this like brilliant guy that was now kind of like, like kept out of that space because he couldn't string together a sentence like in the right way. And, you know, it was just like, obviously horrible. He had to go kind of change his career path. But then like AI comes out and he sees Jasper and he tries it. And he's like, wow, this feels like I can do this again. And he is probably our most creative, unique customer that comes up with like the coolest stuff all the time. And it's like, he just needed these tools to go and participate in that. And obviously he has like, it's a

• 36:14Speaker 01

  unique story around it. But like, think of all the people in the world that just like, can't write. Very well, but like could be great at like marketing campaigns and know people super well. And it's like now they might be like the best marketers in the world, even though they could participate in this. And then, you know, before. And so I think there's just like a whole bunch of things that we're going to unlock there. And again, maybe I'm a really great software engineer, you know, I could build really great apps. Now, I'm not, I'm not a developer, so I actually can't participate in that. But if I could just click a few buttons and do that, like I might be able to come up with, you know, way cooler stuff than my CTO. Now he's great, but uh, but yeah, it's like, you know, since like, I think we're just going to like see a lot more of that. Or it's like people build this across the whole stack regardless of skill sets.

• 36:55Jacob Effron?

  Super interesting. Amazing. Well, Dave, this has been a fascinating conversation. We always like to kind of end our, our podcasts with what we call our quickfire round where we ask a standard set of closing questions and get your take. And so maybe to kick that off, you know, obviously there's, there's a lot of buzz and hype in AI in general right now. Curious kind of of all the things that are talked about, you know, one area that you think's overhyped and one that you think's underhyped.

• 37:16Speaker 01

  You know, I, I actually don't think much is overhyped right now. Maybe one thing that's overhyped is this idea that like, we're all going to be out of work. And, you know, what about the jobs and all of that? Like, I just don't think you ever see that in history at all. Here we are, working like dogs, you know, after the industrial revolution and the internet and all these things. We're supposed to like, you know, put us all to business like, you know, it'd be kind of nice to, to be able to sit on the beach a little bit more. And so maybe that'd be a good thing, but I, I think that's probably like not an issue for a long, long time. And I think underhyped, I just, I think it's still not that useful. Like if I called my parents right now and asked like, hey, have you used AI in your like day to day?

• 37:57Speaker 01

  Like, there's like no chance like they've used AI like somewhere, you know, throughout their day. And so I just think like, there's still so much it's got to do to kind of like cross the chasm and leap into like everyone's everyday life and be truly useful. You've got a lot of really great demos, you know, going viral on Twitter and you got a lot of cool videos, but it's like, where is it? You know, kind of throughout the day, like we're going to, I think we're close. I think like, we're not too far away from that, but I don't think people realize like how impactful that's really going to be when it goes from just, you know, Twitter demos to wow. This is like saving me five hours a day, you know, across my whole life.

• 38:35Speaker 02

  Yeah, it's funny, though, when putting things in perspective with your parents, I actually just had my parents out from South Carolina. And they're so far, they're retired, they're so far removed from this. And we were showing them ChatGPT the other night. And their minds were blown, but I still think they didn't fully understand, despite us trying to explain it, like the power of what's happening and how this is going to change the world. It's always good to get out of our, our relative bubble, right? And connect with the rest of the world and see where those folks are now. Um, another rapid fire question, outside of Jasper, what company or project do you think is especially exciting?

• 39:12Speaker 01

  You know, I think, I think the companies, I've seen kind of a few doing this, but I still think right now generative AI is, is generating content, obviously. And I think what it's gonna do, what it was particularly good at is understanding, you know, natural language and then doing something about it. And I think we're kind of entering this phase where it's like, okay, like now we can like go do things in the real world for me. It can help me book a flight, it can help me figure out, like find that home on Zillow that I'm like searching for without me knowing like the exact way to use all the filters and all of that. It can maybe, you know, go send an email for me. And so I think like, it, it's starting to connect, you know, multiple tools together, which is like super exciting. And like, maybe what I'd be really excited for is this entering like the physical world.

• 39:55Speaker 01

  And obviously like this all this kind of constrained to my computer screen or my phone. And since I turned that off, like, you know, it's not like running the microwave for me or, you know, kind of doing, helping me teach the boys how to throw a football. I don't know, like, but like, if this can kind of enter like the physical world, like that's going to be really exciting instead of it just being a digital product.

• 40:13Jacob Effron?

  Amazing. No, it's a, it's certainly you're seeing early, early signs of that, but I, I think, you know, certainly no, no shortage of a, a potential interesting future applications. You know, David, this has been fascinating. I'm sure folks are going to want to learn more about you, about Jasper. What's the best way for them to kind of do that?

• 40:27Speaker 01

  Yeah, Jasper.ai or from time to time I get inspired and fire out some hot takes on Twitter. So we can go look me up there.

• 40:35Jacob Effron?

  You do have some good Twitter hot takes.

• 40:36Speaker 01

  I go in spurts, man. I try to be like the person that does it and I, you know, again, this is what Jasper solves, but it's like, you know, I just, I don't know, I get busy and I just like, don't want to be like a totally public persona. So I, I'll have like 10 tweets in like two days and then I'll like, be like a month without saying anything.

• 40:55Jacob Effron?

  Amazing.