Beyond The Prompt" shines the spotlight on David Okuniev, the visionary co-founder behind Typeform, as he takes us through the labyrinth of innovation at Typeform Labs. Dive into David's eclectic journey from musician to design maestro and entrepreneurial spirit, which paved the way for Typeform's creation. In this episode, David reveals the ethos driving Typeform Labs, where emerging ideas blossom and new technologies are not just tested but pushed to their limits. Get an insider’s look at the orchestration of international ventures, the spark of radical innovation, and the future of interaction as we know it. Tune in to explore how David's leadership, grounded in authenticity and creativity, disrupts the status quo and shapes the next frontier in AI and product design.
Typeform AI Beta: https://www.typeform.com/ai-features/
📜 Read the transcript for this episode: Transcript of David Okuniev, Founder Typeform: "Redefining User Experience in the Age of AI" |
[00:00:00] Henrik Werdelin: Hey there. Welcome back to the latest episode of beyond the prompt. Where we deep dive into people and their stories about using AI to make the organization a little bit more efficient. I'm your host Henrik Berlin joined by my cohost, Jeremy oddly, today we're thrilled to bring you a true visionary in the realm of user experience. David the co-founder of Typeform.
David is a really incredibly entrepreneurial and as many other successful founders is now running a little incubator, a little research lab. Uh, next to his company where he explores new technology and how it can make his existing product even better.
We explore the frontiers of innovation, the challenges of entrepreneurship and how AI is becoming an integrated part of the future. Let's get started.
[00:00:42] David Okuniev: hey, happy to be here. ~Well, ~just to give you a quick intro on myself, I'm David Okunyev. I'm co founder of Typeform. I run Typeform Labs at Typeform, which I've been running for the last four years now.
~Uh, ~we basically work outside of the main product to bring new emerging ideas to market,~ uh,~ experiment with new technologies and basically trying to push the boundaries of form interaction.~ hilarious. I was blacking out there for a second.~
[00:01:08] Jeremy Utley: ~I'm back. We can't hide anything on Zoom, can we? We can't hide anything. So, David, ~thanks for joining us~ I think, as I, as I told you, uh, Henrik and I are both,~ uh, we're helpless romantics when it comes to enterprise innovation and our desire are because of our constant objective is how do we help folks? In organizations,~ uh,~ discover new and valuable applications of technology.
And of course, generative AI is a radical opportunity for that. And,~ uh,~ ~as, ~as I chatted with you the other day, I thought, wow, our audience would love to hear from you. ~And. It'll just be, let's, let's have a wide ranging conversation. We can pause, go backwards, forwards, you know, anything, anything that you want.~
~Um, we generally have guests introduce themselves at the very end, so we won't start that way, uh, or we won't use this, but ~maybe just for kind of context, you want to give Henrik a little bit of your background.
[00:01:40] David Okuniev: Yeah. ~So, uh, ~born in Belgium. ~Uh, ~my, my mom's British, so I've got a slightly British accent and kind of started my career as a musician and then went into,~ uh,~ design ended up being, ~you know, ~running my own small company doing web design and interface design for startups.
And then eventually started. Typeform,~ uh,~ when I moved ~to, ~to Barcelona,~ like, uh, about, well, it started~ around 12 years ago, ~but moved to Barcelona around 17 years ago. And yeah, I'm just, I love,~ I'm a builder. I love to build stuff. I design,~ I,~ ~I, ~I code and,~ um,~ basically a Typeform today. I run a small group called Typeform Labs.
It's me ~and, ~and,~ uh,~ Through other developers and,~ uh,~ and who's, who does ~kind of ~like business analyst support and everything else,~ and uh, and yeah, just, just like building and, you know, been doing.~ I was the CEO of Typeform for some years, but obviously that wasn't ~my, ~my best calling. So I quit that ~like ~four, five years ago ~and, ~and just, yeah, I've just been building multiple products since then at Typeform in, in Typeform Labs ~and so yeah, I've got a lot of experience with That's incredible.~
~Oh, thank you. So as I got a lot of recent experience with just trying to innovate inside a company that, that struggles to actually innovate, ~
[00:02:29] Henrik Werdelin: ~um, I mean, like, I think your story is very similar, the same thing, built a company about 12 years ago and took it public in 21 and, and kind of like now have to kind of like be in the painful role of trying to get your organization to move at the speed of startup that you remember it used to do, but it doesn't cause it's big.~
~And. You know, like, it has all the same problems that other people are complaining about that Is how to kind of like accept when it's your own. Um, so, uh, yeah, your story sounds very, uh, ~
[00:02:29] David Okuniev: ~familiar.~ I'm also the founder that's been way too long at his,~ uh,~ his own company. ~Uh, so~ I'm like hanging on to ~like, yeah, like ~if I go, like the whole culture goes ~like, ~or whatever's left of it,~ uh,~ but at the same time I'm having fun, I've got like a, ~you know, ~a remit to do.
~Uh, ~to do, you know, what I love to do, although it's not fraught with problems, which I'm now discovering now, which could be an interesting angle as well, trying to align what an innovation team does and trying to align with the rest of the company.
[00:02:53] Jeremy Utley: ~Well, ~let's, why don't we start with that a little bit?
Cause that's pretty interesting. Why has your structure taken shape the way that it has? ~I mean, ~this is, it's an interesting case of you ~kind of, ~you, in a sense, you start from scratch, not because type forms a small organization, but because of your position as the founders, the CEO, you can, you established a structure that you felt worked for you.
We can talk about whether it did or not, but. Why the structure of a lab as a separate entity? Tell us about your thinking there and what you've seen to be some of the benefits and maybe drawbacks of that approach.
[00:03:22] David Okuniev: ~Well, ~I think a lot of the answer ~is, ~is just what you said, right? It's ~like, ~obviously I'm this, I was a CEO, so I created a space for myself, which is, ~you know, ~I want to continue in type form and I wanted to get back to product.
I don't want to lead the main product. I want to be able to innovate. There was a bunch of things that. As CEO, actually it was Co CEO, sorry. So as Co CEO, I wanted to get stuff inside, often inside the roadmap. And obviously there's some tiers of the organization to actually get that implemented. ~You know, ~I worked on several hackathons to do stuff, which would sometimes get inside, inside product roadmap, but I had a bunch of ideas, which I wanted to experiment with.
~So, ~so as I stepped outside, what am I going to do next? ~Well, ~I want to experiment with video in terms of video data collection. So yeah, I, I built a small team. around me,~ uh,~ completely separate to the rest of the organization,~ uh,~ with, without any, it doesn't report into anything. ~It's just, ~it's just independent.
~Um, ~and,~ uh,~ we built, yeah, video ask,~ uh,~ which is a tool that's still going now after three and a half years. ~Uh, ~so that's, that, that was in initial kind of impetus ~of the, ~of the whole thing. ~Um, but you know, um, I'm also a believer that, well, believer, I've learned, I've come to believe, or~ my belief confirmed that radical innovation is ~really, ~really difficult to do inside your own product, right?
~You can, ~you can iterate and you can do some bits of innovation, but ~kind of ~jumping out the box and completely like challenging everything is very difficult if you're like completely inside the box. So you ~kind of like ~have to jump out of it to do other things. So just for example, video, I was just trying to crowbar that into type form.
Initially we could have done it, but it would have,~ uh,~ first of all, we would have to be struggling with the product and all its dependencies. ~Um, ~So we didn't want that, but also it would constrain in terms of what we could do, right, of how far we could imagine this going,~ right,~ or it would all have to align the whole time with typeform.
So yeah, just made sense to,~ to,~ to take it apart. Do
[00:05:10] Henrik Werdelin: you,~ um,~ do you normally, when you come up with something new, do you just follow kind of interestingness? Or do you have more kind of like, here are different areas that it's going to be important for us, unfortunately,
[00:05:27] David Okuniev: yes, I just fall into that trap, ~I guess, ~but I don't know, ~you know, I mean, ~one interesting thing right now is.
~Uh, ~because I'm building this new product called Formless, which we ~kind of ~built to disrupt ourselves. ~Well, ~at least that was ~my, ~my intention is ~like, ~what if we built Typeform with AI, what would it look like? So the kind of the problems I'm having now a little bit, we're trying to line it with the rest of organization is like, Hey, what problem does it solve?
~What is the, ~what is the problem space? ~Well, uh, Uh, ~we're just trying to defend ourselves from being disrupted on our own product with this. And so getting people to understand that when you layer on all that sort of more traditional product thinking, you've got to find a problem. You've got to interview people, blah, blah, which I don't want to say there's no value in that.
It's just my path and how I've done things, including Typeform has never been,~ uh,~ like that. ~So, ~
[00:06:16] Henrik Werdelin: so yeah, I very much agree with that. ~I mean, like, ~and I think. There's a book that I'm obsessing about these days called Why Greatness Can't Be Planned. ~Um, ~it's by a professor in AI,~ uh,~ ~he, ~he was at Florida University and then he left to,~ uh,~ he started a company and it got sold to Uber and he became like the head, like running Uber, the AI labs.
But his point,~ um,~ and the book is really about like, how do you get to artificial general intelligence? And his thesis is very much that instead of trying to be rule based and goal based in your exploration. ~Um, ~you have to be,~ uh,~ follow something he called open endedness, which is associated with this thing he called interestingness, which obviously for an entrepreneur very much resonate because you ~kind of like, ~that's how you operate that you find something that's exciting.
And then you ~kind of ~just ~kind of like ~go that way. But I think he presents a pretty thorough framework for at least within AI code. That's the way that you probably need to ~kind of ~approach it in order to come up with something amazing. And I would argue at entrepreneurship, it's the same.
[00:07:14] David Okuniev: I resonate with that because ~I, ~I've said it before, thought it before, like a kind of product builds itself in a way right when you're doing when you're operating in that very early stage, or you have an idea and then you start putting you're excited you put it together you start putting it in front of some users they feedback ~and and ~and you ~kind of ~shape it and then ~like ~six months down the road it's ~like ~evolved and you're like oh wow we ended up here.
So you ~kind of ~have to let yourself be a bit free roaming to ~kind of ~get where the product needs to go a little bit. And if you have all these constraints, then it's a bit more difficult. One interesting thing, and I haven't read the book,~ but,~ but I know of the, what's called the,~ um,~ the innovators dilemma, right?
Which is, ~you know, ~how can companies innovate because there's this dilemma we have, ~you know, ~we've got this product, it's all working, but we have to protect ourselves from, ~you know, ~competitors actually like leapfrogging us by using other technologies. ~And, ~and what the book goes, if you're familiar with the book, I haven't read it again,~ but,~ but, ~you know, ~just reading ~the, the, ~the main points in it,~ uh, the, ~the recommendation is that ~you, ~you need to build this out.
You need to ~kind of ~act like your competitors. Right inside your own organization. So that's ~kind of ~maps to what I'm doing inform us now pretty well. And that's something that I think big organizations like struggle to understand, right? Because they want the answers to everything. Everything needs to be like planned and how's this going to work?
How's this going to integrate, blah, blah, blah. ~Um, ~
[00:08:34] Jeremy Utley: I wonder if we could start there, David, because it's a great premise and that's something Henrik and I have talked a lot about is this question of if we could rebuild our product in the world of AI, what would it be? How would it work? ~Right? ~And I'm not sure ~that ~that's a fit for every organization and maybe every organization either shouldn't ask it or they're not in a position to commission an exploration around it.
But taking those aside for a second, perhaps we can come back to those questions. Say you as an organization agree It's a worthwhile consideration. What would we do if we could start over from a I and we and who's going to build something that's a I for first that disrupts us. If you just take that the ability to ask the question and the resources required to call it commission the exploration ~once you ~once you're there, what happens next?
And maybe even not theoretically, but walk us through the example of type form given that you started there. What were some of the initial actions that you took in order to discover the answer to the question?
[00:09:28] David Okuniev: ~Okay. Um, ~so I think you're asking me like, how did it initially start? Yeah.
[00:09:36] Jeremy Utley: ~If you've, if you've got this,~ if you get the permission to answer the question, what would our product be if it were AI first?
Yeah. How did I
[00:09:43] David Okuniev: go about?
[00:09:43] Jeremy Utley: Yeah. ~What's, what are your next, like, like ~walk us through the first hundred days, so to speak, ~what do you start? ~
[00:09:46] David Okuniev: ~Let me just talk you through the, through the story that maybe that helps. And maybe you can get the nuggets ~
[00:09:46] Jeremy Utley: ~and we will, we will certainly interrupt you along the way for as friendly.~
~Uh, we're, we're incapable of holding back our enthusiasm. ~
[00:09:46] David Okuniev: ~Look, um,~ ~we were, ~we were building in Typhoon Labs, another product called Relaid where we were experimenting with AI for meetings. ~Um, ~and ~you know, ~we were, We ~kind of ~built this product that built like an, it was an async and live product all on a timeline where we would stitch meetings and async communication together on a timeline.
And then we would summarize the whole thing with an AI and it was actually quite good, but I think we didn't have a chance, like we realized ~after, ~after too late, we wouldn't have a chance against ~the, the, ~the bigger players. And, ~you know, ~everything was just moving too fast. So actually when GPT four came out, ~um.~
~It was, ~it was pretty clear to me that the tech had got to a point where you can actually have a form conversation ~on, on,~ on GPT, actually on GP, GPT chat is actually what kind of spurred it on actually one, one of our investors actually tried to build a prompt to actually ~like, ~Hey, what if a type form could,~ uh,~ talk like that?
So we ~kind of ~confluenced on, on, on the thinking, and then we thought, ~well.~ Well, my first thing was, okay, Relay isn't working, this could be an opportunity that we should explore as soon as possible, because if we don't, if we don't, then other people will. And then there was a consideration, can we build this inside Typeform as such?
And my answer was very quickly, no, for two reasons. One being, ~one,~ the structure of how everything works at Typeform in terms of data collection is just a completely different paradigm. ~Uh, ~and trying to ~map that, ~map that in the renderer and in the data is just, ~you know, unthink,~ was unthinkable at the time.
Now it's a bit more clear, like we understand the whole space and we know how to maybe fit it back in. But at the time I just thought,~ well,~ if we want to move fast on this, we've just got to, we've just got to get to work on it. And,~ uh,~ ~and, and, ~and we just set sail. ~We, ~we built a builder and well, the editor, and we built a renderer and we built a parser to basically.
Well, the trick with formless is not like a bot where you just ~like, ~you record the conversation. We have to extract from the conversation, the very things that the user programs the AI to ask for. So we have a parser and then we extract that data into structured data so people can use it. They can map it to a webhook or to,~ uh, ~ ~you know, ~to,~ uh,~ they can download the responses or they can use Zapier to ~like ~send it to ~any, any, ~any service.
So it's a full end. It's a full end form product. The difference being is that with Typeform, you,~ uh,~ ~the, ~the conversations are predetermined and programmed. When this, it can go in any direction and you can ask questions back for which we have to build all this training, the training capability. So you can.
Put your website address and it can pull in all the information, vectorize it so it's available in the form conversation as well. ~So, ~all these things we just knew we can build,~ uh, inside,~ inside the main product and we just moved ~too, ~too slow. In fact, we don't even want to touch the main product because it's just too many dependencies, too many people to talk with and ~we just, ~we just wouldn't move.
[00:12:33] Jeremy Utley: And do you think about a totally different user? ~So, ~it's one thing to say, What would we be if we were AI first? Is this, and you talk about not wanting to touch the main product, I presume because, ~you know, ~current customers are happy, et cetera, et cetera. ~Is, ~is the new product for someone else, or how do you think about moving existing customers?
[00:12:51] David Okuniev: That wasn't The intention, it was just, hey, what was, like I said, what would this look like? What's, or even easier to understand, the future of forms is this. We've got to build it. How do we build it as quickly as possible? We've got to do it outside the product. So it's almost ~like, yeah,~ now we have a problem.
Who's this for? What's the problem space, et cetera, et cetera. Which is ~kind of ~where we're at now, which is we're one month away ~from, ~from releasing. And we're having this confluence with the rest of the organization is like. How do we sell this? If it crosses over with the same, with potentially the same customers, what use case does it do better?
And the truth is we're still discovering that because we only just launched beta like a couple of months ago and, ~you know, ~it's still, ~you know, ~a low amount of users and we have to talk to a bunch of people. So the truth is, we just don't know these things need like. For me, the innovation cycle, you need like between six months or one year to build, then you have to do a release and you have to test pricing.
I think you really know if you've got product market fit, ~you know, ~maybe it's too late after pricing, but I think in this case it will be after pricing and launching it properly somewhere next year, we'll know ~like ~whether it's got enough traction to,~ to,~ to push forward,~ uh,~ how it. So it's different customers compared to type form.
What's the opportunity and what is the clear problem space that it can actually solve better than type form? Or maybe even it can solve everything better than type from, because people want to be able to have ~high, ~high frequency conversations, as opposed to just fitting things in, they might want to ask questions back and try to get more done, like at that point of interaction, at least maybe that's the way the creator of the form might think about it~ when you had a quiet.~
[00:14:21] Jeremy Utley: ~No, no, go ahead. Henry, please. ~
[00:14:21] Henrik Werdelin: You had a quite hyped, kind of like a beta release of it. At least I remember kind of signing up and ~kind of like~ getting excited and suddenly it was rolling out. ~Um, ~was there anything in how you did that release that you would have redone or is that kind of like the model for trying to ~kind of ~launch these new type products?
[00:14:40] David Okuniev: Yeah, I mean, the first page we just put a form, a formless form ~on ~on the landing page and that was the pre beta and that did ~really, ~really well. The problem ~is, ~is that we took too long to actually get to beta. So by the time we started sending out invites, you know, a lot of that traffic and ~just ~just died off.
So we're ~kind of ~building that back up now. But, ~you know, ~the approach was just Yeah, show the product. front and center and get people to interact with it as soon as possible. ~I ~
[00:15:06] Jeremy Utley: ~think~ it'd be interesting to go backwards a little bit. You, in just ~kind of ~the story you told after Relaid, you said an investor had gotten into chat GPT and said, what if a type form is start having a conversation about type form, right?
When did generative AI really get on your radar as a setting aside kind of gen AI? When did the technology get on your radar? And how did you personally, David, start playing and exploring before? And is there any separation between ~what can, ~what can Gen AI do for our company? And what can Gen AI do?
Period. And just talk about that process in your own
[00:15:46] David Okuniev: experience. So actually we started playing in Typeform Labs ~with, ~with,~ uh,~ with GPT 2,~ uh,~ back in the days of VideoASK. ~Um, ~we built this thing called the Human Chatbot. Well, I don't know if you know VideoASK, but you can basically create a video thread, which also, the question is in video and you can also get responses in video,~ uh,~ or multiple choice.
But we built this thing, Human Chatbot, where you would just Talk to the video and then it would go to the, to a different video. But basically based on what the user was saying, we'd match it to some entities that the user would create. And we used GPT,~ uh,~ at the time to do that. And it worked like actually not too bad.
And we were also using,~ um,~ I think assembly AI to do all that stuff as well, which is a transcription,~ uh,~ life transcription service. So we've been playing with it for a while. Obviously we're playing with it ~with, with, with, ~with relayed. ~Uh, ~but it wasn't, I guess it was never, I think on formless, it's really the main feature, right?
It's ~like, ~Hey, these are AI forms. We weren't saying on relayed or videos, these are AI,~ um,~ These are, this is AI video forms or AI meetings. So it's really now that we've really embraced it. ~Um, ~but yeah, there's, we've been doing it for a while. And then your other question was,~ Well, ~
[00:16:57] Jeremy Utley: ~just, ~just for you, what we've noticed is ~there's, ~there's a, there's something of a, where many organizations struggle, they're asking the question, what can Gen AI do for our business?
And what we've observed over and over again is, You can't answer that question if you can't answer the question of what can Jenny I do for the business. What can it do? And really, a lot of times you actually have to start with some very personal kinds of experiences where you go. It can do that. ~Well, ~it can do that.
And then that's actually what stimulates the imagination. Go ~Well, ~what if we could do this? And I was curious to know if you had a similar kind of a experience yourself and what that was like.
[00:17:34] David Okuniev: Not really, but I would say, I mean, you know, this is different levels of immersion into AI, right? And ~it's, ~it's how much time you spend with it and actually playing ~with, ~with the models as well.
So a lot of people come at this with ~like, ~Hey, we need to get AI on our product because everyone's doing it and everyone's slapping AI features left, right, and center. And I think what I've noticed, and I'm even noticing this through fullness is that. new users are really like the, it's like people have to get used to like working with AI right now, right?
It's easy for images because the connection is very direct, but informants, it's ~like, ~Hey, the AI is going to talk for me. It's going to ask questions. Sometimes it might even hallucinate. Like you have to kind of like shift as a user of this product, you have to shift your mindset to actually. Yeah. ~You know, ~forgive a little bit of the imperfections of AI.
And that is ~kind of ~a problem for some people. So everyone's tipping tiptoeing towards these features, not quite knowing if they're bulletproof or if they're really going to work for them. And I think that's why, ~you know, ~A lot of companies include Typeform,~ uh,~ doing this,~ uh,~ launching AI features. Some of them are getting adopted, but ~some, ~some people are a bit less, a bit more nervous.
For example, I would say on formulas, some people will not want the AI to write the questions for them. They just will go to Typeform and say,~ well,~ I just want to have control here. So it's, I think it's about how much control you want to give over. ~Um, not sure if that answers your questions. Not from one, but.~
[00:18:58] Henrik Werdelin: ~I think there's, ~ I think actually we'll double click on, I think what Jeremy was also asking about, but one thing that I think is super fascinating as a product developer ~and that I'd love to hear if you've experienced too is,~~ um, ~~we have this product called autos. com, A U D O S. which basically is a co pilot for entrepreneurs and it helps people build their business plan and make their first website and stuff like that.~
~One of the things. Oh, A U D O S. Okay. It's launching. Yeah, it's launching this week. It's basically, I'll give you this small brief just because we're about to launch it and you might find it fascinating. It's basically, we run an incubator called PreHype where we build a bunch of successful companies. And so over the last 15 years, we've had this very kind of like a bespoke method of how do you identify a problem, like how to identify a customer, how to identify a problem, how do you then kind of come up with a deal for it and how do you test the market and blah, blah, blah.~
~And so we ended up building. Yeah, like a lot. Um, and so when we got access to OpenAI, we were the same thing. How do we take some of our workflows and build autonomous agents to do it? And one of the things we then come up with was basically, can we create this kind of iron man suit for somebody who says they want to build a company?~
~And what they've been doing is that they, uh, um, we hit people up on Instagram. We put them into DM and then we. Come up with an idea with them. We built first a website, then we build a, uh, a, um, um, a, uh, so first built a, a kind of first we come up with an idea, then we build like a seed investment proposal with them.~
~Then we build a website and then we build a type form, uh, using the API where we, that we asked. them that we asked for them to set to other people. And then we use that those people coming through to create a lookalike audience on Instagram, we helped them kind of like drive traffic and then we off pulled them onto notion and blah, blah, blah.~
~So like ~
[00:19:08] David Okuniev: ~many steps of the journey to try and ~
[00:19:08] Henrik Werdelin: ~enhance. Yeah, I'll cut all this out, Jeremy. Uh, but I, but the, um, I love it. I love it. It's, it's great.~~ ~~But the question, the fascinating thing was.~~ ~And I've built many products in my career is that we then ask at one point, the chat bot to say, we would like to ask the customer for money, but ~I didn't specify where in the flow, um, where the flow that they would ask for it~
~flow, uh, for the cash. I lost you for a ~
[00:19:19] Jeremy Utley: ~second. You said you didn't specify where in the flow you asked for it. And then we lost ~
~you. ~
[00:19:19] Henrik Werdelin: ~Yeah. So we didn't, ~we didn't specify when the flow would be. We just said,~ like,~ we'd like you to ~kind of ~ask for 30 at one point. And what's fascinating is that it would put that question in at random, like at different places, assuming you're having some kind of ~like, ~obviously a logic void and what blew my mind was as a product developer, I'm so used to understand like click for click where people are going and now I ~kind of ~have to reinvent that more like a game designer and saying, I might have spawn points different places, but basically the customer ~kind of like, ~might ~kind of like ~wonder their own.
way as you were doing, and you have a little bit of the same problem, right? Because you have this very unstructured format, but it has to end up in this structured and database structure at the end. Have you seen any of those things where you get and ~kind of ~got surprised on ~how, um, Yeah, ~how it behaved.
[00:20:02] David Okuniev: Yeah. So I used to have hair before I started this project.
[00:20:06] Jeremy Utley: That's a common theme, by the way. ~I used to be,~ my hair used to be black, not so gray.
[00:20:10] David Okuniev: ~So, um, I think Henrik's, uh, frozen again. Are you there? I'm here. Yeah. Okay. Just videos frozen. Oh my God. Uh, I mean, just trying to, uh,~ so on formless, ~you can, you, you build,~ you instruct the AI on a, on one day to do something and then open AI, like they tweak the models all the time.
~So, you know, ~whatever output you got one day will be different the next day. And maybe you like added some. custom prompter to make something work and the next day it doesn't work. So it's that ~kind of ~kind of crazy stuff. So when you're trying to debug problems,~ um, ~ ~you know, ~it's ~like ~one day it's working, one day it's not working, and it's back to working.
And then the other thing is that it seems like,~ like,~ Even the answers that OpenAI give, ~the chat GPT,~ sorry, GPT is giving, feels sometimes even cached for some time. So you don't know if you're actually, the changes that you're making in the prompt to actually make the output better, are actually having any effect.
So ~Well, ~that's on the dev side, but I think the customer will experience this as well, that ~like ~one day it will work one way and the other day it just won't hit the mark. It's just, it just comes with the territory but, ~you know, ~as the models get better then this consistency will actually improve. We built, we actually built Formus on 3.
5 because we wanted to get out early. ~Uh, ~and we couldn't, we thought if we had to scale this, it would just be impossible on four. Now we've moved to four turbo, which is a bit more reasonable. ~Uh, and, ~and it's better, right? In terms of, it's more consistent. ~So, you know, ~over time, these things just improved, but.
I'm not too worried about this because we are in that innovation phase anyway. We can make the mistakes and learn, ~you know? Uh, ~the point is that we get,~ we, we, we, ~we don't give up and carry on to a point where this can really potentially take off. Also, the market needs to adjust. ~Well, ~the full market needs to adjust to something new and that will take time as well.
~'cause competitors are coming into the space as well. There's now a competitor called Ask Merlin, which I think. I think it's going to be backed by Sequoia and they're doing like, uh, AI user, uh, interviews. So, you know, there's, there's a bunch of products that are coming out there. Um, actually I probably need you to edit that part 'cause I'm, I'm not supposed to share the information that I can be backed by Sequoia because I, I got that through, so I just caught, um.~
~Yeah, there are competitors out there like ask Merlin that the coming up and because it makes it's obvious like there are several use cases where having a conversation would be much better than actually just, you know, getting people to fill in boxes. And~
[00:21:40] Henrik Werdelin: in your exploration of AI. Have you found any personal use cases that work.
[00:21:48] David Okuniev: Personal use cases for myself. And
[00:21:51] Henrik Werdelin: so, yeah, ~I mean, like, I think,~
~um, yeah, like, uh, as you've been playing with these, uh, these tools,~ have you found something that you use it for just because as an individual, not necessarily as. Is it for something that you build into
[00:21:57] David Okuniev: your product? Beyond, ~beyond~ just using chat GPT to just use it as an aid for work, because, you know, we're in a small team.
So I, I have to do all the copy for everything. ~So, you know, ~getting some help, some inspiration there. So I definitely use that. ~Um. ~Apart from that, I just, yeah, I just use AI through Formless to build interactions and conversations that I'm going to use to interact ~with, ~with people on, on, on the platform.
~So, you know, um,~~ doing support. I built, I built ~
[00:22:24] Henrik Werdelin: ~a prototype~~, ~ I built a prototype of, uh, it answering taxes from my mom. And so I think somebody like you, who's like a special, like an expert in ~kind of like~ questions and answers, I would imagine that by now.
[00:22:38] David Okuniev: Funny enough though. ~I mean, first of all, my, my,~ I can't build by myself.
~My, I,~ ~I, ~I'm really just a, like ~the, the, ~the end of the front end. ~So I'm really just, uh,~ I build all the UI. I work in react. I do all the, ~you know, ~the basic logic, but I'm not there building the APIs ~or, ~or connecting anything to the backend. So I'm really. I'm really useless by myself. ~Uh, ~but yeah, it would be nice to be able to, ~you know, ~one day to ~kind of, you know, ~improve my knowledge and do that too.
But yeah, ~I haven't, ~I haven't scratched any itches on any personal projects. ~Um, ~yeah.
[00:23:05] Jeremy Utley: What,~ uh,~ what do you think about the broader organization? You may be because of, because Typeform labs may be separate. You may not have a point of view, which is fine, but how do you, it's one thing for your. For your team at Typeform Labs to be, call it AI competent and comfortable.
How does the broader organization think about integrating AI, building levels of comfort? ~Is it, is it,~ is there encouragement broadly? Try these things, use these things. What are ~kind of ~the organizational mechanisms for equipping folks and enabling folks, et cetera, et cetera?
[00:23:38] David Okuniev: ~Yeah. Um, ~I think there's a lot of room for improvement, but we do have a roadmap which does touch on, on, on many things that are going to be using AI to,~ to,~ to enhance the experience.
So it's definitely there. ~Um, you know, ~we have ~like, ~Slack group where everyone that's involved across the organization talking about AI and ~we share, ~we share best practices there. ~Uh, ~but I think people are jumping in and out of it because it's not just like a, ~you know, ~it's not just a daily grind on AI and people aren't just interacting with the models on a daily basis like we are.
So I think since ~we are, ~we are doing that, we can share a lot of the learnings in terms of, ~you know. ~Letting people know, for example, hey, careful because we've experienced this with this model ~that, you know, you,~ it may cause problems for you down the line. So I think that's more or less ~the, the, ~the setup.
~Uh, ~there hasn't been like, ~you know, ~formal like trainings or like cycling very formal sessions if that's what ~you're, ~you're asking for. ~But you know, ~I think everyone's just ~kind of ~scrambling to ~kind of ~adjust ~to, ~to the space. Also, because it's just such a big threat to so many companies, including Typeform, right?
~You see ~GPTs coming out and you see even Gemini in that demo, which is, I don't know if you saw it yesterday.
[00:24:45] Jeremy Utley: Unreal. Unreal. It's
[00:24:47] David Okuniev: unreal, but it's also fake because they've edited,~ well,~ it looks like they've edited ~all the, ~all the latency. So it just feels like this beautiful conversation. Conversation experience, but everything is going that way, right?
~You know, ~we are going to interact with computers in a very different way. Once the technology catches up, we're still like trapped in these older paradigms because the tech is not fast enough yet to really, ~you know, ~make us completely move away from it. ~Uh, so, you know, ~just
[00:25:12] Henrik Werdelin: a little bit on that, because obviously ~you, you, you see ~humane who,~ uh,~ ~you know, ~obviously is this new start tracky, ~kind of ~like.
~Uh, ~device,~ uh,~
[00:25:19] David Okuniev: ~yeah, yeah, ~yeah, the pin, ~um, ~
[00:25:19] Henrik Werdelin: ~and, ~and then you have pie and you have a lot of these different LMS now that you can talk to. Meanwhile, humans,~ uh,~ obviously used to talk to humans. And I think what we're chatting, we have areas where we used to chat with people. What I've noticed with a few products I've been involved in is that if you try to put a chat bot on a website.
it can be a little bit awkward for people to use it. But if you put it into a DM or WhatsApp, then yeah, like then people understand it better because they're used to that behavior. ~What do you think is going to be,~ how long are we going to wait until this idea of using conversation and interactive conversation as a way of directing?
The computer more often is something that we would adapt more or ~will we have to adopt like, ~will the computer have to adopt with the behavior that people already have?
[00:26:04] David Okuniev: Yeah, I mean, a few thoughts here. One, like we've had this interaction for a long time with Siri, but no one's kind of taken up and actually the turnaround in terms of coming back and answering.
~I mean, it's, it's, ~it's. It's not a proper conversation, obviously, but, ~you know, ~the turnaround time in terms of latency is not that bad. ~And, ~and no one's screaming like everything needs to be done like Siri, ~like ~10 years ago, whenever it came out. So ~that ~that's,~ um,~ that's one thing to think about. The other thing that I was thinking is that ~I think~ you may have spent a lot of time.
I don't know if they're running their own models, but they've definitely worked very hard ~on, ~on, on the latency issue. And being able, because the problem is you've got to understand and then you've got to produce and stream the audio back as fast as possible, right? And, ~you know, ~unless you own some piece of that tech, it's, ~you know, ~I just speak for ourselves,~ like,~ we're just relying on third party tools.
If we did this, we would use open AI 11 labs and ~just ~just stream that all back. But still, ~you know, ~there's a bit of latency, which just would feel awkward. ~So, ~yeah. I just think when the tools, like back to my original point, when the tools become a little bit more,~ um,~ democratized to really do this well, I think you're going to see,~ uh,~ just a lot more interfaces being less clunky and less awkward, right?
~This, ~this idea of having a chatbot on your website, it's just a, it's just a commodity. I don't know where it's in, but would you really. Would you really talk to it with your voice and then wait for it to come back? ~It's like you just want to,~ you're just on a website. You don't want to make that much effort.
But if it's, if the experience is that seamless and that real time and that human, then I think like suddenly like people will make that switch. ~Right. ~And it will be like,~ well,~ I can't have an intercom chat on my website, whatever it is, I need to have this new thing. ~Right. ~And we're ~kind of ~doing that with formless a little bit.
We're not embracing audio yet because it's not ready yet, but we're trying to like, get away from ~like ~what a chatbot typically looks like, which is like this robot talking with lines of text and the bottom and the text box at the bottom. We're trying to make it more like an on brand experience,~ uh,~ where we bring in images, videos.
The text types out it's big fonts and it's just feels more of a design experience. So I just think we're trying to pave our way towards that eventually. ~But you know, ~back to also, sorry, just to add one more point, but back to that point, that all this stuff may be democratizing the threats to companies such as ourselves.
~Like, ~I don't know what the world looks like when so much of our interfacing with computers is not done through like a website or an app. ~And, and, and, ~and therefore. Yeah. ~How, how, ~how does anyone differentiate themselves if everyone's also using the same tools? So you'd have to think that the big players are going to just suck everything up, right?
I don't know. ~Like if, ~if, ~I mean, ~I just think of Typeform as an example, we really have to try and find ways to protect ourselves for the, so that doesn't happen by a big player that wants to just take our space, like OpenAI has killed like some companies that have been doing what they've been like trying to work with APIs and then now they're going all product orientated because they know their survival can be, it's not gonna be based on the LLMs that they have, it's based on the products that they build on top of their LLMs.
~So, ~so yeah, it's ~uh, ~I don't know if it's a race to the bottom, ~race ~
[00:29:13] Jeremy Utley: ~to the,~ I do think that there's differentiate or there's a, you're enlightened in that you're very sober minded about the fact that you're going to be in the crosshairs soon. ~You know, ~I heard Mike Noop of Zapier recently talked about how.
~You know, ~automation, workflow, automation and AI is basically a convergent zone. Like they knew we have to do something or we're gone, right? I think there are probably other parts of the economy, ~you know, ~outside of the startup ecosystem, Silicon Valley, et cetera, where there's, there could be this kind of,~ uh,~ one could say delusion or at the very least, ~you know, ~reality that,~ uh,~ we're a long way from being affected by this.
Do you have any, how do you think about. Knowing what the horizon of AI impacting a business is like, are there questions that you would suggest people ask to, to have ~a, ~a sober minded estimate for themselves? ~I, ~
[00:30:00] David Okuniev: I think first, I think you need to act like as if it is a big threat, even if for the back of your mind, I can agree with people saying that,~ like,~ Hey, not, ~you know, ~nothing's going to change like that quickly.
It's going to take time, but I think if you. Play that game early. And I think you have an advantage anyway. ~Um, and I lost the thread of your last, the question that you were asking. How ~
[00:30:17] Jeremy Utley: ~do you, how do you begin to assess the extent of the threat? Yeah. And just how to think about it. ~
[00:30:17] David Okuniev: ~It's different.~ It's really difficult because ~I mean, ~things are just moving so so fast.
And then wherever you think one day it suddenly completely changes or accelerates the next day. So~ I don't know, I think even just looking at Google's demo yesterday made me Like even accelerate my thinking in terms of the original thinking on formless, which would be like the future of forms is going to be fully conversation and fully multimodal.~
~Sorry,~ when I first started forms, I thought fully conversational. Now seeing the Google demo, I'm like, Hey, interaction and data collection comes into that is going to be completely fully modal. There'll be voice, ~there'll be, ~there'll be, ~you know, ~vision, everything will come into one thing. So. ~Yeah,~ you can ask yourself questions one day and then the next day, everything just starts changing.
And then there's a even better LLM that comes out that does something like even better. And then the game changes again. Yeah.
[00:30:52] Jeremy Utley: I like one of the former guests we had last week. One of the things that he said was you need to update the assessment every six months. It's not that you make an assessment at this point in time.
You put a calendar reminder on every six months. Let's. ~Let's, ~let's reassess what's the state of this technology, et cetera. But I
[00:31:09] David Okuniev: think you will naturally reassess all the time ~as you, ~as you see things coming out. ~I mean, ~things are coming out all the time. You just have to go on Twitter, ~you know, ~and if you get enough in your feed about AI, there's just crazy stuff coming out every day.
And it's just moving so, so fast. ~I mean, ~
[00:31:22] Henrik Werdelin: it's ~very, ~very complicated to keep up. ~I mean, like, ~I feel like I spent four or five hours a day just. Researching the latest and I have to have time keeping up. Which is ~kind of ~mind blowing.~ Do you think there's going to be.~ Are there going to be like pauses, or do you think that this is just the
[00:31:35] David Okuniev: new normal?
~Well, ~it should only accelerate now, in theory. ~I mean, ~look, here's a good example,~ like, you know, ~when this whole AI explosion started, and ~you know, ~image and even video creation with Runway,~ um, ~ ~you know, ~everyone's talking about like threats ~to the, ~to the movie industry, and ~soon you'll be able to like,~ everyone was coming out with like wild theories in terms of ~like, ~Hey, you'll be able to just tell the eye ~to build a, ~to build a feature movie and just watch that.
~I mean, ~I remember ~like~ that question, certainly maybe six months ago. And now what you're seeing is started like glimpses of that. ~Right. ~Although six months ago, maybe people were saying, Hey, that's a long way off. And it's, ~you know, ~it's ~kind of ~crazy. Like you start seeing ~like, what, ~what PICA, I don't know, P I K A.
Yeah. Yeah. They're cool. ~And, ~and especially in animation, like obviously you have to compile things together, ~but you know, ~if you like spend enough time and ~like kind of ~worked with the technology, you could get a pretty good result even today. ~Right. ~Question is ~like, ~if. ~I mean, ~they're probably working on this.
It's just, it's the same as that new tool. I can't remember the name starts with them. The,~ uh,~ the upscales, the,~ um,~ ~the, ~the image, right? Invents the rest of the AI in it. ~What's it called? It's like a French word. Well, you've, have you seen all this upscaling stuff where you basically. Right. I presume it is a little distortion.~
~So you run it through this and it just like, wow, just makes it looks incredible.~ So this is ~kind of ~like self healing, but so you could do the same thing with video. It's ~like, ~Hey, look at this, detect all the imperfections and try to ~just, just, just, ~just self heal it until you get, get
[00:32:42] Jeremy Utley: it to a point where you give me the Tarantino version.
Give me the Spielberg version. Give me the Ron Howard version, right? Yeah. But
[00:32:48] David Okuniev: it's just a question of how long it's going to take to get to that fidelity. And when you get to that fidelity ~and it just, You know, ~there's going to be a point where the straw breaks the camel's back, and then people are going to start to create these things.
~And then, yeah,~ and then all these kind of doomsday scenarios that people were saying six months ago about the movie industry could start ~to really, ~to really happen.
[00:33:05] Jeremy Utley: It's like a long way off is a lot sooner than it used
[00:33:07] David Okuniev: to be. ~I think you have, like I said,~ I think you have to act like it is because otherwise you're just going to, because things are moving so fast, you're likely to be caught with your pants down.
It might be already too late. Just, ~you know, ~just getting started is hard. So just stay in the game and ~keep on, ~keep on working.
[00:33:21] Jeremy Utley: Okay. So getting pragmatic, ~I mean, ~especially former CEO, you, I think you, you have a good answer to this question. If you start from the premise that it's going to happen, how do you carve off resources to do it?
~To attend to it? Meaning how do you, what,~ ~what, ~what is the, what are the financial structures to make sure that you're able to do the AI exploration before it's too late?
[00:33:47] David Okuniev: ~Well, it's just all I can talk is about our story because that's all that's all I've been in. Um, ~so I believe if I have to give it to another company, I think you always need people that have the freedom to just completely experiment and just come up with the next ~crazy, ~crazy thing.
I wonder frame of websites, for example,~ uh,~ if that. ~You know, ~who built that? And was this a small team that just went off and say, Hey, find something new and build that, and now it's their main business, right? ~Uh, ~so I think you need to have to put some resources and investment into having a team that can just operate without any constraints, but you.
~You know, ~your current product is going to be your bread and butter. So you also will want to be incorporating,~ uh,~ AI features into it. And I don't think ~it's a, ~it's a financial, it's not, it'll cost what it'll cost. It's more about what's the roadmap investment, right? What are the things that you're going to trade off to actually make those decisions?
And obviously every company is going to be different. I think it just has to come down to ~what, ~what things are really going to enhance the user experience,~ uh,~ or really help people in type form. ~You know, ~now you can start with AI. So ~it will, ~it will actually write all the, although the questions are predetermined, it will actually write all the questions.
So you'll say, Hey, I want a form that does ~this, this, ~this, blah, blah, blah. And we'll just build the type form for you. So those things like, ~you know, ~those high value things you need to like, Carve space in roadmap to build. ~Um, ~yeah, I'm not sure if that's the answer you're looking for, but that's probably the best what I
[00:35:06] Jeremy Utley: can.
~I think, ~I think a lot of organizations go, we're already so budget constrained. We're already cutting, ~you know, ~it's, and you say, wow, we need to devote a couple of people like David of David's salary level and expertise to, to just go explore. Yeah. ~I think, ~I think it is more rare. That ever before for organizations to have capacity to explore.
[00:35:27] David Okuniev: But you say that which when you think what kind of organization do you have in mind like a really big organization like a sort of, ~you know, ~200 people startup or what do you have in mind ~when you when you think about that? I think once the company, it's, you know, just putting a group of people onto a problem.~
~They do that all the time, right? ~
[00:35:37] Jeremy Utley: ~You'd be surprised. ~
[00:35:37] David Okuniev: ~Okay, well, you know more about it than I do, obviously.~
[00:35:37] Jeremy Utley: I think that's mine and Henrik's kind of perennial challenges. It's, we want innovation, we want it now, and we don't have any money for it. ~You know, ~and ~it's ~it's some version of that. I need it. It's not we can't take time.
We can't run experiments. If we do, none of them can fail. And we don't really have any budget to afford it. You go, Okay,~ well,~ that's a really small sandbox. ~Right? ~
[00:36:00] David Okuniev: ~Yeah, ~yeah. ~Well, ~I don't know. Obviously, that's a problem, right? ~I mean, Product can't just stay,~ product offering can't just stay static forever.
And at some point you're going to have to, ~you know, ~not be, especially if you're a leader in your space, ~you need to, ~you need to lead. So you need to invent the next things that your product can do. And so the best way to do that is to have a group of people that think outside the box and you continue in best time.
Otherwise. ~You know, ~someone else is going to do it and you'll just become the incumbent. And,~ um,~ ~I mean, ~it's just a story of startups, right? It's ~like, ~I don't know that cheesy phrase innovate or die. So it's surprising to hear that you're seeing so many companies,~ um,~ following that pattern, right? ~Um, again, I think ~
[00:36:35] Henrik Werdelin: ~one of the issues that the other companies,~ I think one of the issues the other companies Don't have that you do is they have a founder that is still part of the innovation process.
And I think the way that you think as a, I think one of the, some of the properties of a founder is that of agitation, that of ~kind of ~like panic, but it's also of the ability of atomizing a problem and then using technology to try to solve it in a new way. ~And, ~and so one of the things that. I'm curious about is ~S.~
S. A. I give people I am in suits. And so you have somebody who is very capable and with a very deep inside, for example, to what a customer want now when they now have the ability of producing much more. ~Well, ~that changed the type of people that we have in organizations because you, in essence, would be able to have more people kind of work less mini founders.
And if that is true, how do we find and develop those? Because obviously it's not a job spec you can write. It's ~kind of ~like an attitude more than necessarily a
[00:37:35] David Okuniev: skill set. Yeah, ~actually, you've kind of, ~yeah, you've hit upon something which, yeah, I recognize a type form. We've tried to do some kind of.
Bigger radical ideas inside, inside the product teams. They're usually led by, by product managers and I've seen a couple of projects just completely fail because they were so ~kind of ~just working outside of everything else and just trying to get something of value over the line. What I observed is that, ~you know, ~people aren't willing to take many risks or try things and they're always ~kind of ~like ~reporting up, ~reporting up to leadership and saying, Hey, this is what we've done.
Are you okay with this? They don't have that. You really need that. ~Uh, you need to, ~you need to be able to operate ~like ~as if you're building a new company. So it's that founder mentality. So if you want to do that within your own, let's say product organization, ~you have to, ~you have to have people not just have that mindset.
You also have explicitly say is ~like, ~you're in charge,~ like,~ just go off and do this. I don't, ~you know, ~just move as fast as you can. So I think,~ uh, I think it's, it's, it's mine.~ It's the right mindset, obviously being a founder and obviously having that remit in the company, coming down from being CEO, like ~it was, it was, ~it was easy.
So there's a good fit. And now. So my question is ~why, ~why do organizations find it so hard to actually innovate? Maybe it's not budget. It's just the confidence and investing in the people that can actually make it happen and not being a whole waste of time and just some people playing with toys in the corner.
~So, so yeah,~ you need that. You need pioneers. Basically, ~and ~
[00:39:04] Jeremy Utley: ~if you think, I mean,~ ultimately, I mean, it's, we're coming back to the beginning of the conversation, ~you know, ~why is a lab required, et cetera. And ~what, ~what you observe is that folks are drawn to an organization and certainty and safety for some reasons, and the kind of people that are drawn to an established role with a salary ~with, you know, ~They're just a different constitution.
It's not better or worse, right? But then when you talk about the founder's mindset, a fixed salary for a known objective and responsibilities does not result in a founder's mindset. It doesn't attract the founder's mindset,
[00:39:37] David Okuniev: right? Yeah. And actually we've brought people into the team from Typeform, which haven't panned out, for example.
~Uh, ~actually most of the people hired into Typeform, that actually came from outside, some from inside the work, but clearly ~there were, ~there were people that joined the team and it just, yeah, they just didn't have that mindset, that whole ambiguity of, ~you know, ~having, ~you know, ~to have everything clear and, ~you know.~
Working on shaky ground ~just, ~just didn't, they couldn't operate. And ~so, so, ~so yeah, ~it's a, ~it's a people thing. Yeah, definitely. ~Well, ~and
[00:40:06] Jeremy Utley: I think going back to the conversation we had, ~I mean, I love, I, I just messaged Henrik privately. I love that you, you can't even fathom not innovating, but that's a function of your mindset, right? ~
Your mindset is innovation. First, I think if you were to build
[00:40:13] David Okuniev: things and getting excited about stuff, like I wouldn't be at Typeform if I couldn't do that, I'm not there to continue to have a salary, right? I'm there to build,~ uh, you know, ~I like designing, I like coding and, ~you know, ~I don't want to be playing with toys on the side either.
I want to bring value to the company, right? And I want to be successful and learnings and so forth. And if I can't do that, then there's no, I shouldn't be in the company. But I
[00:40:36] Jeremy Utley: would say ~if, ~if you take that as call it X, okay, that characteristic, call it X over Y being the total population of the organization.
It's in the vast majority of organizations. It's an exceedingly small, likely sub one percentage. There's ~very, ~very few people who have that attitude. ~Right. ~And in some, I think there's,~ it's,~ it's approaching zero. There's just, there really aren't people who go, it's,~ uh, ~ ~I'm, ~I'm willing to embrace the ambiguity.
I'm willing to, I'm not doing this because of a salary, right? The five things you just rattled off are exceptional characteristics.
[00:41:10] David Okuniev: ~Well, I mean, ~likewise,~ if,~ ~if, ~if I had to work in the main organization, I would completely fail like completely. It just wouldn't work. ~Right. ~It never will achieve product officer.
~Like, ~forget about it. Like it just would never work. I need to work in a small team, very close to people and just hit the ground. ~Um, so yeah, yeah. ~~It's a people thing. Um, but you know, that you were. So just one more thing. There's a lot of people like that out there, but they generally maybe don't want to join companies, right?~
~They want to do their own thing. So ~
[00:41:25] Jeremy Utley: ~that's actually and it's your point. It's not a better or worse I mean, I love what you said about I couldn't be a chief product officers, right? That's a very valuable useful role to be performed in an organization. You go. It's not me You know And I think that's where we have to be careful not to place value judgments on particular characteristics or types that all are needed It's just clear there there where we need to be clear is Some mindsets aren't appropriate to this, the innovation objective, other mindsets, the innovation mindset isn't appropriate to the CPL objective, for example.~
~So ~
[00:41:25] David Okuniev: ~in that first hand, I can, uh, I will testify in court to that. ~
[00:41:25] Henrik Werdelin: ~If you were to,~ if you were to offer an advice to somebody who sat in an organization and they had something, or everything you say really resonate, but they sit in the co organization, they're not in the labs. What's their path to ~kind of like ~scratch the itch of doing something entrepreneurial?
Is that to just email the CEO? Is that is to find the founder? Is it just to build the product on the side and come and show it? ~Like ~what's your. What's your two cents on like somebody listening and going ~like, ~Hey, I'd like to do more of
[00:41:53] David Okuniev: that. It hasn't really happened when people could, we do these things called type jams every end of every product cycle, which are five weeks and people are doing like exceptionally cool things, but they're doing it all inside the product.
So the point of these like kind of hackathons is to find ideas like into the product, but seemingly no one is. It's ~kind of ~working outside of the bounds. So it's interesting. ~Like, ~it hasn't really come up that much. ~Um, maybe they just think that it's not going to get, like you said, it might not get like picked up or they don't know how to, how to, how to pick it up.~
~I mean, ~
[00:42:20] Henrik Werdelin: ~they want to make it for their own. Yeah. All that ~
[00:42:20] David Okuniev: ~institutionalized Typeform as a process, maybe it's like, Hey, got an innovative idea. Like we have an incubator. We don't have an incubator ideas. We have a team, which, which, uh, sorry,~ we don't incubate innovation team. We just have an innovation team that incubates ideas.
So it's
[00:42:26] Jeremy Utley: right. They're synonymous. Yeah. Yeah. Yeah. Mechanism is the people. It's not a mecca. ~It's, ~it's actually the people
[00:42:31] David Okuniev: themselves. Yeah, exactly. But, ~you know, I don't know, like~ our chief pro officer was like talking to me about this,~ like,~ could we like incubate. Other like small startups in type form doing something under our vision.
Thing is, it's really hard to ~like ~pull off and ~like. ~Myself personally, if I'm fully focused on a product, just trying to do that well, it's just ~very, ~very hard. So unless you have ~like ~a chief in someone like this in chief innovation officer kind of status, but then you're back into that whole kind of corporate game again, and maybe that will kill innovation again.
So ~Right. ~
[00:43:01] Jeremy Utley: ~Don't get us~ now. Anytime the phrase chief innovation officer gets mentioned, we have to end the show. So that's the end. very much for being here. We really appreciate it.
[00:43:09] David Okuniev: ~Amazing. ~
[00:43:09] Jeremy Utley: ~Boom. Drop the mic, my friend. ~
[00:43:09] Henrik Werdelin: ~Anything that we forgot to ask you that we should include before we let you go? ~
[00:43:09] Jeremy Utley: ~David, would you? Okay, now let's go to the very beginning. Would you be so kind as to introduce yourself to our audience? If we usually take a couple takes.~
~So just, you know, spit something out. We can give you feedback or whatever. Uh, think a very, think 15 25, ~
[00:43:09] Henrik Werdelin: ~4. 4 45 minutes. 45 minutes, ~
[00:43:09] David Okuniev: ~hour. Yeah. I'll just say the whole thing again. Um, hey, my name is David Oev and, uh, I'm co-founder of Typeform. Sorry, the Typeform . ~
[00:43:09] Jeremy Utley: ~Now you got me nervous. What is the company?~
~What is the company? ~
[00:43:09] David Okuniev: ~Never know my own company's name. Hi, this is David Okunyev. Um, wait, hold on. I'll do this as if it's like part of a podcast. ~~Hey, happy to be here. Well, just to give you a quick intro on myself, I'm David Okunyev. I'm co founder of Typeform. I run Typeform Labs at Typeform, which I've been running for the last four years now.~
~Uh, we basically work outside of the main product to bring new emerging ideas to market, uh, experiment with new technologies and basically trying to push the boundaries of form interaction~~. ~~Amazing. ~
[00:43:09] Jeremy Utley: ~That's super cool. No, as a, ~as a happy time typeform customer for years.
And now that you shared formless with me, ~you know, ~a couple months back,~ I,~ I love it. I'm a huge fan and I'm rooting for you. Excited to see,~ um,~ if you want to let folks know perhaps what's a link that they can visit and how do they
[00:43:24] David Okuniev: AI and it's an open beta at the moment, you can just sign up and get started.
It's free. You get. ~Uh, ~you get a number of responses that you can play with,~ uh,~ and then you can upgrade. We give you a nice, during beta, we give you around 250 conversations. So you're good to go. And then you can extend that by paying. ~And, um, when we launched the product, because it's AI, it's going to be, there won't be much of a free plan.~
~We'll, we'll reduce the number of free responses down, sorry, free conversations down, but. You know, we're trying to make it competitive, but you know, it's AI and the costs are just a little bit crazy and not very scalable at the moment. So we're actually going to be running it almost at cost at the beginning until we see the prices going down.~
~So, so yeah, hopefully we can make a dent, uh, in the space and, you know, even disrupt ourselves and take it from there. ~
[00:43:42] Jeremy Utley: ~Amazing.~ Amazing. David, thanks for joining us, especially after your surgery. It's been a really enlightening conversation.
[00:43:47] David Okuniev: ~A lot of people think it's American, so at least in the past they used to. But, uh, look, it's been, it's been good, but it's, it's, it's fraught with difficulties. Like it's, uh, it was much more of a walk in a park in, in the past. Now it's, you know, you know, growing is like much more challenging and the company is going to have to like really diversify more into what it's doing.~
~We kind of like this beautiful no code tool for um, for having Let's say conversations, but they're fake conversations on type form. And that really just like kicked off. We have like 150, 000 customers or more now, I think, but you know, we're trying to edge to like a billion conversations now a year, like of our, our customers having with, with users.~
~So it's not the whole internet, but it's a, it's a little slice. And so, yeah, it's pretty, pretty crazy. I just think we came at the right time. You know, we're just, you know, no one was looking at forms to try and like put design thinking to it. And we thought, Hey, wouldn't it be cool if a form was like a conversation.~
~And so, so we built Typeform and now we really can, you know, make a form like a conversation, so not just like a conversation, we can make a for me conversation, which is right. For me, it's like going back to the beginning. I'm formless. So that's amazing. ~
[00:43:47] Jeremy Utley: ~That's super cool. No, as a, as a happy time typeform customer for years.~
~And now that you shared formless with me, you know, a couple months back, I, I love it. I'm a huge fan and I'm rooting for you. Excited to see, um, if you want to let folks know perhaps what's a link that they can visit and how do they ~
[00:43:47] David Okuniev: ~AI and it's an open beta at the moment, you can just sign up and get started.~
~It's free. You get. Uh, you get a number of responses that you can play with, uh, and then you can upgrade. We give you a nice, during beta, we give you around 250 conversations. So you're good to go. And then you can extend that by paying. And, um, when we launched the product, because it's AI, it's going to be, there won't be much of a free plan.~
~We'll, we'll reduce the number of free responses down, sorry, free conversations down, but. You know, we're trying to make it competitive, but you know, it's AI and the costs are just a little bit crazy and not very scalable at the moment. So we're actually going to be running it almost at cost at the beginning until we see the prices going down.~
~So, so yeah, hopefully we can make a dent, uh, in the space and, you know, even disrupt ourselves and take it from there. ~
[00:43:47] Jeremy Utley: ~Amazing. Amazing. David, thanks for joining us, especially after your surgery. It's been a really enlightening conversation. ~
[00:43:47] David Okuniev: ~Now people can have context. It sounds like I might have had like a bypass.~
[00:43:47] Jeremy Utley: ~Exactly. I really, I really hope you you treat the condition. ~
[00:43:47] Henrik Werdelin: ~We hope to, we hope to talk to you again in the future.~
[00:43:47] David Okuniev: ~Uh, do you seem a little bit familiar in one way? Uh, I don't know if I'm mixing up with some, but I heard someone talk that like, seems quite similar to you at a conference, a 0. 9 capital conference in, in San Francisco. I think like maybe like eight years ago. Do 0. 9 capital or, or even know them? I mean, ~
[00:43:47] Henrik Werdelin: ~I, I have been, I think so.~
~It seems familiar to me. I don't have anything with them, but I have been very. Formiscus on the speaking, uh, circuit. And so you always had nine years ago. Would you do ~
[00:43:47] David Okuniev: ~you always wear black? Yeah, this person was in black as well. And I don't know. I think he was Danish as well, which you are Danish. ~
[00:43:47] Henrik Werdelin: ~No, I mean, it was a Danish person in black that says something interesting.~
~Then definitely me. It was somebody says, then I'm just not sure ~
[00:43:47] David Okuniev: ~that it's a long time ago. So I don't know, but, uh, and this person wasn't a 0. 9 investment. I think that person just was like a guest speaker. ~
[00:43:47] Henrik Werdelin: ~So then it seems very likely ~
[00:43:47] David Okuniev: ~maybe. Yeah, ~
[00:43:47] Jeremy Utley: ~he was dashing, but he's, but he's only more charismatic now.~
~Right? Yeah, ~
[00:43:47] David Okuniev: ~maybe, you know, Christoph Jans. Well, maybe you just don't remember. I ~
[00:43:47] Henrik Werdelin: ~think so, but I also don't remember any names. I have like this thing in my brain that I have no name recall at all, and it's very embarrassing. To the point where when Bark started, you know, and we are like, you know, thousands of people now.~
~But when we started, we were like four people. Of course, you know, my two co founders and then we had two staff, just people working. And one of them was this incredible woman that ran our engineering team, but I could never remember what she was called because I can't remember names. And so we would do like our Wednesday all hands meeting five people.~
~And I would be like, and now, uh, engineering. What kind of idiot are you? I didn't know. I know now she's called Becky, but, uh, but ~
[00:43:47] David Okuniev: ~I, I have the same, sometimes I can't remember the names of people in my leadership team, so don't worry. Like I, I always get stuck. I don't know why. It's like, I don't know. I'm a bit dyslexic, I think.~
~Hey, ~
[00:43:47] Jeremy Utley: ~you, hey, you people respond to that. It's good. Yeah. ~
[00:43:47] David Okuniev: ~No, the name comes, but sometimes you just get blocked as like, you know the name. It's like, oh, just, you just go blank. ~
[00:43:47] Jeremy Utley: ~That's so funny. Awesome. ~
[00:43:47] David Okuniev: ~Probably some plug with everything. Thank ~
[00:43:47] Jeremy Utley: ~you. Thank you, David. We'll, we'll, um, we'll, we'll clean it up and we'll, uh, we'll send you, you can listen to make sure we got Sequoia sufficiently redacted and we'll probably be posting early next year.~
~Probably February. So just FYI. ~
[00:43:47] David Okuniev: ~All right. Well, good speaking to you and hope this was helpful and amazing. ~
[00:43:47] Jeremy Utley: ~Amazing. Thank you. ~
[00:43:47] David Okuniev: ~Great. Thanks a lot. ~
[00:43:47] Jeremy Utley: ~Take care. Adios.~
[00:43:47] Henrik Werdelin: ~We're here. Jeremy. ~
[00:43:47] Jeremy Utley: ~What is that? That's cool. Yeah, what a cool guy, huh? ~
[00:43:47] Henrik Werdelin: ~I love that we're doing this, like, this is incredible, huh? Like, just listening to all these people ~
[00:43:47] Jeremy Utley: ~and Learning, learning. ~
[00:43:47] Henrik Werdelin: ~Getting You look very dapper today. Are you going to You're about You're Are you getting christened? Am I getting what?~
~Christian, you just look like you're, look like you're a church outfit. Oh, no, that's funny. ~
[00:43:47] Jeremy Utley: ~Oh, no, no, no. I've got a, I've got a class I teach in Tokyo a few times a year and the, um, last session's today. So I try to, I try to look, I try to look the part of a professor. Cause everybody in this class wear suits, like all these Japanese folks wear suits, you know, ~
[00:43:47] Henrik Werdelin: ~they're, we need to go there.~
~Do you know anybody cool over there? If we want to launch autos in Japan, have any, uh, people that, ~
[00:43:47] Jeremy Utley: ~you know, you, you saying that is not an answer to your question. I met the most famous designer in Japan. Uh, now to Fukusawa, he hosted me at his house when I was there in January, he hosted me for like, he literally built everything.~
~And Rick, I'm talking the sink, the chairs. The everything he built, and he has this incredible prototyping studio in his basement. He's a, he's a world famous designer. I mean, I read an article in Monocle about him, and I just reached out just through his form, like his website form. Hey, Stanford professor, come to town.~
~He, his assistant wrote me back. We'd love to host you. Please come. So anyway, I had a couple of hours with him. I don't know if he's doing anything with AI, but he could be a fun person to have on the podcast. Um, To answer your question about knowing people, I know, yeah, I mean, I know a bunch of folks in Japan, but they're all kind of corporate y types.~
~Um, yeah. There's one, there's one interesting VC in Japan who I could introduce you to who might be interesting. ~
[00:43:47] Henrik Werdelin: [TAKE AWAY SECTION] ~Let's go back to the podcast. What, what was,~ Jeremy.
[00:43:50] Jeremy Utley: ~What is that?~ That's cool. Yeah, what a cool guy, huh?
[00:43:53] Henrik Werdelin: I love that we're doing this,~ like,~ this is incredible, huh? ~Like, ~just listening to all these people
[00:43:57] Jeremy Utley: and Learning, learning.
[00:43:59] Henrik Werdelin: what stood out to you from this conversation we just had?
[00:44:02] Jeremy Utley: ~Yeah. Let me look at my notes real quick. Um,~
~I mean, obviously~ I think one of the big things is that question, what if we could have built it with AI? ~You know, what if, ~what if we're starting today and we were trying to do the thing that we do, but with AI, what would we build? I think it's a great, it's a foundational question that every business should be asking.
[00:44:20] Henrik Werdelin: I 100 percent agree with that. I also find it just fascinating when you have somebody as innovative as him, ~as clearly so, achieved SAP. How complicated it is for him, and honestly, for somebody like myself,~ to drive innovation when the organization becomes big. And I think in many ways, it ~kind of ~talks a little bit to it being a structural problem, more than necessarily like a talent, like abilities.
Problem, ~right? You know, there are, you know, I met many innovation officers or CEOs that just kind of like, I think in many ways of putting about themselves that they are not able to get the organization to move faster. And the reality is that when you have a company that is building a cold business, and we're really like the business is to make sure that the bread and butter is being made every day, and then suddenly kind of like creating the space and creating kind of like the.~
Yeah, the process to do something completely new, it's just incredibly complicated. And it's ~kind of ~very validating to hear somebody like him who started to build an amazing product and keep doing so articulate how difficult to find it too.
[00:44:51] Jeremy Utley: ~Well, ~and it's wild, ~right? Because most, I would say~ most established organizations would see Typeform as a startup of sorts and they go, Oh, surely a startup can move quickly.
And yet the founder and former CEO goes. Really quickly, the product roadmap becomes so clear that we can't just put new stuff in it. And if I want to keep innovating, if I want to think about who's going to disrupt us, I actually have to have a totally separate structure to do that. I love the reference to the innovators dilemma and how he's thought about resolving it and basically building a team to effectively red team themselves and say, who's going to put us out of business?
What are they going to be trying? I think that he does. He has a mindset that the future is inevitable. And that disruption is inevitable ~that I think there's many folks tend to lack, they tend to think the status quo is inevitable. The status quo will persist. And I think his kind of his foundational starting point is things are changing.~
Things will continue to change. What will we do given that things will change rather than believing that things are going to remain the same.
[00:45:38] Henrik Werdelin: ~I think that's, I think we'll do the outro after that, after that, um, ~
[00:45:38] Jeremy Utley: ~right. I also, I also, by the way, I love the phrases just like as, you know, pull phrases, whatever. Resolving the innovators dilemma is interesting. Uh, and then also the future is inevitable. I think is a super interesting phrase, you know, because it's like, it's a tautology, you know, I love tautology, but it's.~
~But just so true, the future is inevitable. Um, ~
[00:45:38] Henrik Werdelin: ~I have on a practical note, I'm almost done with Greg's interview. I think I will have this done a little bit of luck today, but otherwise kind of Monday, Tuesday, then we have the first thing to go out. Everything else looks pretty nice. I have the whole thing set up as you've seen.~
~I made the website. Uh, I'll be able to post to these other ones. I think I'll also, as you suggested, then see if I can cut a YouTube version, even it's going to be janky and like, because obviously we edited it and so people will jump. I don't think it matters because I think people who watch it, they are just listening to it as much as they watch it.~
~Um, so. ~
[00:45:38] Jeremy Utley: ~You can do one of two things with YouTube. You can do one of two things. You can either, you can just post the audio and just have kind of like the, the cover be the visual, or you can actually post the video, um, and have YouTube be kind of like the visual. I think Tim Ferriss posts a video right of his interviews.~
~Whereas like I met a guy yesterday, he's got a, you know, GSB podcast is very successful. I was looking at their YouTube. It's just the cover art. They don't, there's no video. Okay. So I think you probably do either. Yeah, and then the question is kind of what makes it clickbaity, right? Because YouTube's all, I know one of the reasons that I started pushing to YouTube is because, or I've kind of started to explore, I wouldn't say it's been, I'm not like there yet, but one reason I want to is because YouTube's pushing podcasts.~
~And they're investing resources and investing support in podcasts. So, you know, ~
[00:45:38] Henrik Werdelin: ~what I probably do is I'll do the still, and then I'll try to see if I can find one or two video thing that I'll try to, I haven't played with it that I will upload as shorts. I, somebody was telling me through the entrepreneur.~
~For no great wine that because shorts is a new platform, you can get quite a lot of traction there if you've built something for that. And so why don't, I mean like the, you know, Greg's kind of like the, uh, AI is testosterone from, from the brains. Yeah. So like that, you know, I'll get some of those in. Yeah, ~
[00:45:38] Jeremy Utley: ~that'd be cool.~
~Yeah, definitely. That's great. Right on. And then you're happy. I'll, ~
[00:45:38] Henrik Werdelin: ~I'll, I'll send you, ~
[00:45:38] Jeremy Utley: ~I was just going to say, do we want to block a time next week to record the JJ thing? Cause I don't think we have a JJ outro right now, right? Do we want to do that? That's ~
[00:45:38] Henrik Werdelin: ~true. Yeah. Do you want to put a fine time right now?~
[00:45:38] Jeremy Utley: ~Yeah. Why not? Let's see.~
[00:45:38] Henrik Werdelin: ~Monday, Tuesday, I'm sort of out, so Wednesday would work for me. ~
[00:45:38] Jeremy Utley: ~Yeah. Wednesday is fine by me. I can do in the, in the morning. ~
[00:45:38] Henrik Werdelin: ~Let me see. My wife is out for dinner. 11 too late? ~
[00:45:38] Jeremy Utley: ~11 is not too late. Let's do 11. That's fine. I mean, whatever works for you. Is that what works best for you? ~
[00:45:38] Henrik Werdelin: ~I just need to figure out.~
~It says my wife is for dinner. So I imagine that I'm then responsible for getting the kids to bed. And so. Perfect. ~
[00:45:38] Jeremy Utley: ~Let's do 11. That's great. I'll drop you a calendar invite. ~
[00:45:38] Henrik Werdelin: ~We can move it. And worst case, I'll just move it around. Yeah. ~
[00:45:38] Jeremy Utley: ~Yeah. No worries. Cool. All right, my friend. Enjoy New York. Have fun. How long are you in town?~
~All right. I'm flying out tomorrow. ~
[00:45:38] Henrik Werdelin: ~I've been like in kind of just been crazy. Like I, I've eaten too much and drank too much and now I flying overnight and then I'm going straight to Switzerland. So it's like a very, I did get a book on a very big. Uh, talk in May in Switzerland, which is one of those, there'll be filmed, very professional.~
~So maybe I need to pick your brain on like doing a keynote, ~
[00:45:38] Jeremy Utley: ~but yeah, it sounds, Hey, also one, one other idea I had, you were mentioning, you mentioned that book, why greatness can't be planned. And it sounds like he's in AI too. I haven't read the book yet, but I read your post and I liked it. We should invite him to the show.~
[00:45:38] Henrik Werdelin: ~Yeah. I already did. Like I recorded his interview with him, uh, and almost thought about it and I should cut it into. One. I don't think he has that much Gen AI perspective though. Okay. ~
[00:45:38] Jeremy Utley: ~Okay. Okay. Then no worries. I just thought he could be cool. Yeah, that's fine. I mean, you ~
[00:45:38] Henrik Werdelin: ~should, you should, you should meet with him.~
~He's San Francisco. Oh, cool. I found him to be like super ~
[00:45:38] Jeremy Utley: ~fascinating. Huh? Yeah. Love to. Yeah. Introduce me. That'd be cool. I'd be happy to. And I don't think, at this point, do we have any more interviews scheduled? I think we probably need to get on the scheduling front. Yeah, ~
[00:45:38] Henrik Werdelin: ~I, I feel I, I emailed a few on, uh, on Twitter and some of them was like, yeah, sounds good.~
~And then I was like, can we go over email and then they dropped it. So I'll try to push a little bit there, ~
[00:45:38] Jeremy Utley: ~but we got to figure out what the, uh, conversate just like Instagram message or like, what's the conversational way to get somebody all the way to booking in the platform. Where we make contact, right?~
~That way you don't have to change it. ~
[00:45:38] Henrik Werdelin: ~I think also will help to have the podcast actually up and running, right? Like ~
[00:45:38] Jeremy Utley: ~you can point to it. Yeah, totally. Totally. Okay, cool. Right on. Okay, buddy. Have a great day. All right. Bye.~
That totally makes sense. And with that. That's all for today's episode of beyond the prompt. If you liked it just a little bit would really appreciate if you'd share the episode and maybe go in and rate it and like it. Uh, on your favorite podcast platform until next time. Be well.