Latest / STR Unpacked / Ten Blue Links Are Dead: What AI Really Means for How Guests Find You with Milena Nikolić Heywa Labs

Transcript

Ten Blue Links Are Dead: What AI Really Means for How Guests Find You with Milena Nikolić Heywa Labs · STR Unpacked

• Ben: Hello and welcome to Short Term Rentals Unpacked. So my guest today has never run a short term rental in her life. She probably couldn't tell you what a channel manager is or what it feels like to get a message about a broken boiler at 2am. But she might be the most useful guest I've ever had on the show. So Milena was the CTO of Trainline. She was the engineering director at Google Play and building things used by billions of people. And she's just raised a seed round for her new company, Heywa Labs, which is building way all of us are going to be using AI in a couple of years from now. So here's why she's on. Most of us in the industry know that AI exists, but I personally believe that almost none of us are actually using it well. ⁓ this episode is about closing that gap. By the end, you'll understand what this stuff really is, where heading. and one shift in how guests are starting to behave that I reckon should be keeping every operator up at night. Milena's never run a rental, but she is one of those rare people who can explain anything, really, no matter how complicated in a way that we can understand, right? So that's all I'm doing today is aiming Milena's knowledge at our world. Milena, thank you so much for joining me. Milena: Absolute pleasure of being here, ⁓ Ben and then thank thanks for such kind words. Ben: No worries, it's a genuine pleasure. So I think just before we get into everything, I think that the story of actually how you ended up where you are is so interesting to me. So you're from rural Serbia, right? So how did you go from rural Serbia to CTO of Trainline in London? Milena: Indeed, yes. Yeah, yeah. Yeah, born and raised. gosh, for okay. ⁓ takes me back a few years. ⁓ I think I was ⁓ so I I I like to say that I was engineer, you know, my entire life, or I was software engineer my entire life. you know, to to be interpreted because, you when when I was very young, they just meant spending a lot of time playing with with computers. and then, you know, I got into programming I don't know, I was probably twelve and ⁓ Ben: Yeah. Milena: And sort of went into high school with that had kind of a lot of subjects related to that, and then studied computer science at in in in Nis Serbia ⁓ related to it. ⁓ And ⁓ sometime the end of uni, ⁓ I and got ⁓ internship from Google, and they offered any city in ⁓ for me to to take that internship, kind of the summer internship in. ⁓ And I chose London because it sounded like the most exciting city, and I always wanted to visit. Ben: Obviously. Milena: ⁓ came, fell in love with London and decided to kind of come back and live here. Got a full-time offer ⁓ from Google there straight away and joined as a new grad software engineer back in gosh 2008. Yeah, quite a few years ago. and yeah, kind of ⁓ been in London now for 18 years. Here is home very much. I've stayed at Google for a very, very long time, you know, kind of Ben: No. Milena: I think I was there for close to thirteen years. ⁓ it's a fantastic company. It's kind of ⁓ it makes it very hard to leave once you join in. They took really good care of me. I kind of got promoted or had a new project every year or two, so it was very interesting. Didn't feel really like I was in the in the same ⁓ place for too long. I was software engineer most of that time from again new grad levels up until ⁓ some sort of you know staff engineer level, whatever I was. And then I moved into engineering leadership roles. So I was engineering manager and then engineering director. And just before I left, I was ⁓ senior director for Google Play Developer Ecosystem. And then after more than a decade, I kind of had this massive itch just to do something else. I was kind of like, ⁓ my god, I can't believe I've spent so much time in one single company and kind of this was not how My career was supposed to go. ⁓ so then kind of started looking around. you know, we're very fortunate in London that there is lots of choice in the in in the tech industry. ⁓ I really wanted somewhere smaller, which I guess compared to Google almost anywhere is. I wanted somewhere where I felt that I could, you know, based on my own contributions, ⁓ Ben: Yeah. Milena: I can really see the impact on the company. Whereas at Google, even as a moderately senior person, ⁓ you know, I was still one of the thousand directors. And it just kind of like I really wanted to to see how the actual companies run end-to-end. And ⁓ always loved trains, ⁓ I loved consumer products. So kind of when conversations started happening with train line, that was kind of quite, quite compelling. ⁓ I like the brand, I like the product ⁓ that it builds, and I felt there was a lot of potential in the business to turn what was I guess essentially ⁓ ⁓ like a really well known and successful British brand into broader ⁓ european market. So yeah kind of fled Google back in 20 gosh twenty-one I think and ⁓ joined train line spent four years there as chief technology officer fantastic times had had lots of fun Ben: Mm-hmm. Milena: learned a lot from kind of yeah, kind of other other executives and then sort of ⁓ at some stage ⁓ just had too much a little bit too much of an itch to to to start a company of my own, which I can imagine you you you know what I'm talking about and probably many, many of the people listening to podcasts. So yeah, that was kind of ⁓ that's how that change happens as well. Ben: Mm-hmm. Yeah. And how did that feel going from, I guess, know, when you're your CTO of a company that so many people have heard of to leaving to into the great unknown abyss of starting your own company with a much smaller team? Like, how did that feel your side? Milena: Yeah yeah yeah. I mean it's ⁓ it was exciting. I think more than anything it was exciting. I think I had the itch to start something of my own for the longest of time. ⁓ what I didn't have, I think, was the confidence. I didn't feel like I felt that Google prepared me extremely well for, you know, to build technology, to build ⁓ certain, you know, certainly kind of software but also products that people love. But I knew nothing else about the business. I didn't know, you know, how to run a PL. I know how to market the product. I didn't know how to think about anything else kind of beyond beyond the core ⁓ software and the and the product side of it. And I feel that's something that I got a lot of exposure at, you know, kind of at at at a train line. And I was focusing on on kind of on purpose as well, just making sure they actually understand how the how the business is is ran. And then yeah, a few years of that and I kind of had a look and I was like, you know what? Ben: Mm-hmm. Milena: I think I now know enough that I I have the courage to do it. So it was sort of once I had that confidence, then it was sort of there there's still obviously a little bit of fear and uncertainty and you know, gosh, is this crazy? Should I be doing this? Is this the right time? And it's never the right time. It's never the right time to do something that's kind of disruptive like that. ⁓ but I felt ⁓ it was very interesting time in the industry because of AI, and I'm sure we'll talk more about it. Ben: Yeah. Milena: kind of I think in my personal life as well it felt like a right time and kind of like, hey, I don't think I can wait another ten years to to to kind of to do something like this. So it's sort of like it felt like quite a few things fell into place. I was just incredibly excited to to to be doing it, very committed, you know, kind of with enough confidence that I could genuinely believe and not lie to myself that I think I can I can build something very exciting and successful. So yeah Ben: Mm-hmm. Amazing. And so you jumped two feet into the world of AI. Milena: ⁓ yeah. ⁓ yeah. Yeah, that was the again like very interesting time. It was obvious pretty quickly, I think, back in twenty gosh, what was it now? Late twenty twenty three or no, maybe yeah, late twenty twenty two or s sometime in twenty twenty three, that kind of when Chat GPT sort of burst into the scene and kind of captured so much of our imagination and and and time and ⁓ kind of attention. That things will be changing. And you know, kind of like I think it's there are many, many a fad, you know, many fads come and go. And kind of I think like three to six months into this, I think it was getting pretty obvious that this is actually one of those big changes that are here to stay. The other one that I kind of I was always comparing it with different things that I could remember in my lifetime and career. I don't fully really remember in detail that transition from pre-internet times to internet times. You know, I was small enough not not to quite have that perspective. Ben: Yeah. Milena: But I very clearly remember pre-mobile and post-mobile times. You know, that was early in my career. You know, I was very fortunate at the time to have been on Android team at Google and Google Play, or what is now Google Play team, and see that go from you know zero users to three billion users and zero money to you know billions in revenue and kind of all that. So sort of, and I saw how the rest of the ecosystem, how the entire world had to adapt from Thinking about websites only, to thinking about mobile apps and and kind of how all of us as you as users, as consumers, had to adapt from you know. not having to having this thing in our hands all the time and doing so much through it and kind of all that. And it just very quickly with AI it became obvious that it's kind of like that. That this is gonna be the change that's gonna be quite profound in terms of if possibly even more profound than the mobile shift was, in terms of how we run our businesses, how we spend our time, how we kind of engage with information or getting things done and kind of all that. And yeah just I felt like n now is the time to be building. It's kind of, you know, there will be a lot of change, it'll be exciting, but it's sort of ⁓ it felt like possibly the rewards and opportunities will be disproportionate because so much is changing. Ben: Yeah, that makes sense. ⁓ we get into everything that Heywa Labs do, ⁓ I'm really excited to get into, ⁓ would say just kind of bring it back to industry. think AI ⁓ is very interesting conversation and it's a very live conversation at the moment. And I think what I see in operators that I talk to is ⁓ they from some might use ChatGDP to help write their listings to some are basically building their entire apps on Claude. ⁓ So was wondering just for kind of... ⁓ the people who are aware of AI but maybe not aware of what it is or how it works. Are you able to give like a little bit of an explainer to what AI essentially is? Milena: Yeah, yeah, yeah. so the term is, you know, very old. It's probably from, you know, sixties or s or or or seventies. And I think artificial intelligence just as a term refers to machines, computers being able to exercise behaviors we attribute to ⁓ generally to humans. So kind of something that goes beyond computation and beyond fast automation of tasks into something that's more like judgment or creativity or critical thinking or you know kind of having an opinion or kind of something like that that goes or ⁓ that that goes beyond just kind of computing or calculating or optimizing tasks and kind of this this other set of things. So that's that's just kind of general idea for like if machines can do that, then we have artificial intelligence. Now that Bar has massively moved over the years and centuries, and things we thought were AI now are just, you know, kind of we think are actually solved and pretty easy and simple. So I think that's also there is no like one single easy definition. We kind of keep evolving what this is. I think the term of the moment is AGI, like artificial general and general intelligence. So not just having a computer be able to do one of these tasks like creative writing or solving hard mathematical problems or, you know, kind of ⁓ whatever, ⁓ on its own, but having Computer have a bill ⁓ have a comprehensive coverage of different areas in a similar way in which a human could. So all of us to to a certain degree can exercise critical thinking and and and judgment in terms of task, all of us can exercise ⁓ creativity and and kind of various different things. So kind of, and we're now talking like when does it get to this point where machines can kind of do all of those things to a similar degree or or better than than than we can do? And then the other part of that is like when can they start self-improving? When when can kind of computer software become so smart that it can improve itself? And that then kind of that gets to that that point of kind of artificial general intelligence. And then the current kind of actual implementation technology that Ben: Yeah. Milena: We use to to the kind of the thing we talk about most as AI now are large language models, right? Like that's the kind of and again, like AI is this more vague, general, kind of aspirational term. And LLM is a very specific algorithmical solution for how to do that. And really, what large language models are, ⁓ they are token predictors. So they kind of Given a couple of words, let's say, they really well predict what the next word would be. And in kind of when you just explain it in this in this way, it's incredibly dumb, right? You're kind of like, wait, what? Like this thing just kind of knows how to predict the next thing based on the previous thing. That's is is is it actually possible? Is is it is that creative at all? Is that Ben: Okay. Milena: you know, capable of thinking, is that capable of judgment if that's the only thing it does? It just predicts the next kind of word that comes in based on what you what you put in there. And I think the thing that surprised all of us is actually how how how well that works. Like ultimately the I think that the the big discovery in the last couple of years was that if you throw all of world's information, so the entirety of the internet, if you ⁓ kind of throw all of that at that algorithm, it turns out that yes Ben: Yeah. Milena: Yes you can you get some of these emergent behaviors like ⁓ Critical thinking and judgment and creative writing and problem solving and reasoning and kind of all of that. You get all of that out of this algorithmical system that has been trained on all of the data that we're aware of, like the entirety of the internet, all the books, all the YouTube videos, everything else that we can find, that is enough. Which I think is just interesting in its own, right? Like it means that, you know, all these problems that we think are deeply original and new, they're probably not. There is enough knowledge in the humanity that's been somehow encoded and once Ben: Yeah. Milena: kind of funnel through this algorithm does allow us to solve our most common problems and kind of all that. So that that's kind of I'm hoping that that kind of demystifies this the thing that sounds pretty ⁓ pretty magical is actually a fairly well understood algorithm. I think the magic was finally ⁓ companies being rich enough to be able to throw enough compute power to process all all this data that we have, and now that's producing results which you know ⁓ seem seem s se seem seem kind of ⁓ irrelevant to what we described AI as Ben: Yeah, and I think it obviously is anyone that kind of uses it, it's astounding how good it is. Like you say, like when you lay it out, it is incredible, but it works so well considering that's just how it works. And I think obviously what you're doing with Heywa is taking it to the next level, right? With how people interact with AI itself, right? Milena: Yes, exactly. one of the insights that myself and Martin, my co-founder, had very early on is we sort of we really liked how kind of How how Chat GPT or kind of general agents like Claude or Gemini have changed the way we interact with information online. Like kind of we all used to go to Google and you type in a couple of keywords. So you would say, you know, short term rentals, London, and you maybe find a couple of websites that seem relevant, and then you click on the website and you navigate, you kind of play with that website, maybe you click on the the few kind of all that. Now there's this one magic box where you know you just put in way more detail what you're looking for. Like I'm looking For a bright and airy room for two people for any of the weekends in May in Bristol, with you know, and I'm kind of this is that it's it's a very kind of suddenly this thing is way more expressive, understands us, and instead of us hopping over multiple websites to patch up the information and you know, kind of understand what's actually happening, kind of there is this one place where it's all magically happening. So, like it was very immediately obvious to us that. there is like a interface paradigm shift happening because of this. Now we we we we were also shocked that kind of that in ultimately the way that interface ⁓ change it is playing out is the conversational chat bot. I don't think that's a good UI. You know, I think we have like decades of digital media and print media before that that have taught us that people learn better and consume information better and engage with it if it has pictures, if it's structured in a nicer way, if it's formatted, if you know you have larger things and you know, you have tables and blocks and diagrams and kind of ⁓ different things. And the fact that this beautifully colorful and emotionally evocative internet that we had for the last, you know. 30 years is now suddenly reduced to walls and walls of black text on white surface. You know, it's kind of like that was so my me and my co-founder were like, this kind of doesn't make sense. Suddenly we have this incre extremely powerful tool in front of us, and instead of building better experiences, we have regressed, regressed into just like text in, text out, you know, type lots of text, get lots of text, type lots of text. So kind of that was ultimately the thesis we wanted to explore within Haywell Labs, like ⁓ how. ⁓ Using all this fantastic new technical capability that we have, how does the way we interact with the with information, with services, with websites, with apps, with however you call that, how does that change? And we wanted that to be ⁓ visual, we wanted it to be structured, we wanted to be the most user friendly, well designed thing that it can be. And we also wanted it to be brand aligned, and you know, kind of there is so much that companies and brands communicate. Communicate through the way they look and present information and tone of voice and how they talk, that you can never summarize in ⁓ you know, Claude or Chat GPT summary. And that's where kind of like we felt there is a big gap both for businesses and for consumers, in kind of ⁓ getting the output, like being able to very briefly specify what you want in natural language with all the expressiveness that natural language gives you, that filters on the website will never be able to give you in the same way. But then what you get back is structured, beautiful UX, ⁓ something that really works, that you can navigate easily on your phone, that you can tap, and so on. Ben: That's amazing. I mean, it makes sense, isn't it? Like, if you look at all of the apps that most people use, they're all picture based, like Instagram, TikTok, others that I can't think of in this moment, but they're all they're all like that. Okay, that makes a lot of sense. Milena: Yes. Yeah. Or or or even going into going into your industry, like and I'm you know, I'm hoping this will resonate with with kind of people on the on the on the pod, is ⁓ you know, there is a reason why ⁓ you kind of the the the all all the kind of accommodation ⁓ services have evolved to have the UX which allows you to specify the things and to book them in the same way. There is a reason why they have pictures because people want to see, you know, kind of where you're gonna stay. There is a reason why you have to specify obviously the dates and the and the times and kind of all that and there are certain restrictions I think that come from the rigidity of the input that we have right now with with filters. And I think some of that can be fixed with natural input. But People will still wanna see the options. They will wanna be able to compare them, they'll be wanna be able to decide. And that's not something that fits just into walls of text in ChatGPT, you know, that that will have to look and feel different to in order to work. Ben: So in terms of like actual practicality, if you use Heywa and you say, okay, I'm looking for a two bed in Soho this weekend. So then that will basically bring up the pictures of available options or talk me through how that element will work. Milena: Yeah. ⁓ so there are quite a few different things we're thinking about here that that we think are important. We do think it's nice to start with a short natural languaging because it's the most flexible. But from there on, kind of very often that's not quite enough. You might say, I want ⁓ you know, kind of ⁓ a two bed in soho, and probably, you know, kind of ⁓ the system needs to understand to have rough dates in mind, is are you flexible? Is it kind of like so then Ben: Mm-hmm. Mm-hmm. Milena: People not having to write an essay when they interact with a service, I think is gonna be pretty critical going forward. You know, it's kind of I if I'm at my laptop and I'm doing work stuff, I don't mind writing six or seven sentences to specify something really way or really well, like writing a good prompt. But if I'm on my phone, I'm on the go, I'm around, Be typing seven sentences to specify, like, you know, I want to open a service, tap, tap, tap, maybe write a few words and kind of go from there. So I think there is that optimized that entire side around optimizing the input. And in in travel space in general, and you know, I as you know, I work the train line, so I know travel space, certainly on the transport side, maybe more than accommodation. It is technically complex. You have the number of people, you have ⁓ people's r restrictions, sometimes, you know, pages or mobility restrictions, you have dates, you have ticket types, you have the all of these different things. Again, having to write all that out in words is hard work. But having like a nice, ⁓ flexible but user-friendly flow that you can tap through very quickly, I think is a big part of that. And then ultimately you will get some options. And ⁓ being able to go through those options with pictures, with visuals, being able to actually understand the trade-offs, and that's the other thing that I think we get now with with with AI systems. Let's say there is a short list of three flats in in in Soho, and like kind of I want system to summarize how they fit my requirements, my intent, and my restrictions. For example, you can say The first one is cl closest to the tube station, but the second one has nicest restaurants, ⁓ you know, just a minute away. And then this third one is whatever, closest to the museums or or kind of something like that. Like ultimately, we now have the ability to personalize kind of what you get. So rather than getting that grid or list of hundreds and hundreds of hotels or flats or you know, flights or whatever it is, you can get three actually well suited to what What you're trying to do. So then kind of I think the interesting problem becomes just that entire journey. You know, it's kind of you start with a few words and then the system shows you a few options, but also says, ⁓ but actually, just tap here which of these times or dates are kind of interesting for you. So you you kind of very easily select some kind of rough timing on dating, including flexible options like any weekend in May or something like that. Ben: Mm-hmm. Milena: that kind of that selection changes. Then you say, I would like a place that's brighter, or I would like a place that accepts pets, or kind of something like that. And that kind of selection evolves with you. So you can kind of very easily navigate it, you kind of until you until you get to, you know, one or two or three with very clearly outlined ⁓ trade offs. And I think that sort of ⁓ you know that that's the type of solutions we can build now that I don't think we were really able to easily do in the in the kind of in the in the old world. Ben: And that's so interesting. And one thing I wanted to ask you is like one of the things that I think we're seeing as a result of AI is kind of how customer behavior is changing. And you reference, say, you know, like where you used to like go into Google and basically say, I'm looking for two bed in, so, but now it's much more that. people I think are delegating these searches to the AI, right? For it ⁓ to bring options to them a way that you Do you ⁓ Heywa is in a really good position to kind of take that changing in customer behavior or have you found kind of any challenges in that shift to your end? Milena: Yeah. There is a lot of discussion around this. Like in in in general, how are, you know, p brands and and and businesses gonna be affected by this this entire change? You know, there is the sort of you know, there are people who are saying all the aggregators are dead as part of this because now the agent can in the background go and kind of do Ben: Mm-hmm. Milena: Do the job of aggregators instead. I'm not entirely sure that's necessarily how it goes. I still think the core experience of ⁓ you know, w where you find the service and and kind of and and and and what that means and someone actually optimizing to to to to to to kind of to make this really good. mega model companies, you know, Ontropic and OpenAI and and and Google, kind of the the makers of these biggest tools, they're fighting on so many ⁓ battles, they will not be able to go into every single niche and build the most effective experience for every single ⁓ kind of ⁓ thing. So I I still think there is a lot of Opportunity for kind of all businesses out there to find the way they add value. So it's kind of like in that sense, I don't think we will all we might all be starting in ChatGPT or Claude or or Gemini or even our browser. You know, that that magic box can be almost anywhere. And the model might route you to somewhere. But ultimately, I still think the the brand and company then needs to win you over with the service. It needs to win you over with ⁓ ease of booking. with anything else that comes as rest of the of of kind of of of of that business. So it's sort of like, ⁓ I think a lot of different interpretations. Some people are saying these type of businesses will disappear, these type of businesses like I I think everything is still up for grabs. It's unclear, but I wouldn't bet that ⁓ foundational model companies just eat up absolutely the entire stack. Like I I don't quite think it'll go that way because there are a lot of hard problems to solve kind of all along. Ben: Hmm. Yeah, that makes a lot of sense. And also I think as well, like one thing I've always thought when you see like all these industries are no longer going to exist as a result of AI is like, well, that's assuming that industry doesn't change in any way, which of course they will, you know? One conversation that is bubbling up, and I wouldn't say everyone's talking about it. Milena: Exactly. Of course they will, of course they will. Exactly. That's exactly what it is, exactly. Ben: but it's definitely becoming more of like a thing that I'm hearing much more at conferences and in podcasts and everything is at the moment kind of a general flow of how people get bookings is we use OTAs which can be Airbnb and booking.com which is the vast majority of people will book using an OTA along those lines. A lot of people will have their own website but in our world you're pretty lucky if you're getting above 15 % direct bookings. So it's really hard to drive that and it's Milena: Listen yeah. Ben: it's a massive motivation for people to drive to their own websites because you earn twice, right? So one of the conversations is like, okay, can we use AI to actually drive more people directly to our website? Which I think kind of brings me on to kind of two questions for you. And first of all is like, you obviously are creating this business in a way of monetizing it. Milena: Yes, exactly. Ben: Is there going to be a monetization play where someone basically goes, too bad in Soho, I can pay for my property to be there. And is there like an SEO play to make me more discoverable with companies like and maybe others. ⁓ Milena: For sure. ⁓ b both of those. This is like it's kind of ⁓ for for now we're you know focused on growing users, so we're kind of not not actively looking to monetize right away, but later in the year, you know, we hope we have enough critical mass to to for that to be ⁓ you know, worth doing and and kind of feasible. for sure we want to ⁓ Ben: Mm-hmm. Milena: We believe we're building a like a really fun consumer product with a very wide reach. And we do not want to charge for that product. ⁓ so we will need to make money through ads and affiliate revenue. So that's exactly what you described is precisely the monetization path we're we're we're we're looking into. So yes, that'll definitely be be there. ⁓ and then kind of on the SEO side and kind of all that. ⁓ I think there are a lot of interesting signals, and again, this is where ⁓ Ben: Mm-hmm. Milena: We're on the lookout to how the rest of the industry is changing and how all of this is going. Ben: Mm-hmm. Milena: But again, we we really we talk about ourselves as a generative UX company. We like the the kind of quality of the customer experience is at the core of what we do. And we think as we continue to innovate on on on the UX, we think we can instead of the SEO and traditional way in which you plant the right keywords and kind of based on the keywords the search engine decides to rank you higher not, we think we can do something based on the experience as a whole. So if kind of your service supports intent based. Filtering and you know a couple of different things, we think that that would be our version of SEO. So kind of nothing to announce on that today, but we are thinking about that space a lot. And and yeah, we think it's it's it's really important to be to be having both organic and ⁓ support for both organic and paid ⁓ acquisition models for for businesses on the platform, yes. Ben: Mm-hmm. Yeah, exactly. And it makes a lot of sense. And do you think that that, you know, in a few years time, I don't want to be sensationalist and say that will replace Google because obviously, like we just said, Google will change and adjust. But do think that that'll be a much more normal thing that people do with general AI companies? Milena: Yeah. I d I do. Like I mean I kind of like again without sounding too sensationalist, I do think ten blue links are dead. And I think those ad slots at the top that look the same as ten blue links, I I that just feels at this stage so twenty years ago, you know, it's kind of and I just cannot like I think there would be it's very hard for Google and again I have a soft spot for the company because it works so long there, kinda this is that ⁓ so just it's not a but ⁓ and you know, I think they're gr they have to very gradually evolve it because a very large revenue stream. Ben: Yeah. Yeah. Milena: For Google comes from exactly that that part of the business. So it's kind of just scrapping that and starting with something completely new is not something, you know, it's kind of that business Kindle for its shareholders. but I think they will have to evolve it away from that because yeah, that's again, 10 blue links, yeah. I I I don't think that's the future. Ben: Yeah, and do you think in that scenario, like say for example, I put ⁓ a property and I pay for your platform basically to push it. Do you think that is a more... ⁓ How do I phrase this? Like you're going to get a higher success rate using those visuals and using much more of a story than you might do in other forms, even on Airbnb, for example. Do you think it will create a better success rate? Milena: Yeah. I ⁓ cer certainly I think like platform native visual advertising is just proven to always convert better than kind of anything else. I do think our product sits at a very interesting ⁓ intersection between being intent heavy. So people come into Haywan type they type in what they want. It's kind of similar like what they do do with Google. So they would type in, you know, kind of the the the ⁓ short and rentals and so on, kind of ⁓ or something like that. So the moment you know intent, you're not just reading Teaching people called and that converts better. And then, kind of like ultimately, you are your the ad helps them achieve what they're trying to achieve. So, of course, it's gonna convert better. ⁓ and then the more visual it is, that also means that it converts better. So, it's kind of, I think we have those two things that work in our favor, and that's why we're hoping that we'll be able to have an ⁓ ad ecosystem that works for kind of all the parties, and that's kind of ultimately ⁓ you know serves users well and is profitable for us, and is profitable for. for advert ⁓ advertisers as well. Ben: Nice, mean, it's so exciting. I do you... I guess it's, don't know how easy it is to tell this point because it's not that widely used. But do think when we get to the point where people will be paying ⁓ advertising ⁓ Heywa and ChatGDP and Claude and Perplexity and all of these, do you think ⁓ pricing will be kind of level to what we're doing with Google Ads? Or do you think it's going to be a much more of a higher premium to ⁓ in this kind of way? Milena: It's it's pretty tricky to judge. We have done some modeling and it it's also ⁓ kind of it varies by the industry and the goals and the product and kind of all that. So I wouldn't necessarily bet on it. ⁓ our early estimates are that it's probably gonna be more similar to advertising on social media, ⁓ than than kind of on on even though again per vertical that may or may not be similar to to to U UI C and advertising on on on Google. But that's where like kind of ⁓ Ben: Yeah. Okay. Mm-hmm. Milena: Yeah, based on our modelling we think it'll be similar to social media advertising. Ben: Fantastic, fantastic. That's so interesting. And I guess before we came online, you said that one of the things that you were looking for was kind of people to help you out in terms of like looking for beta testers. Milena: Yeah. Yeah, so mm there is something exciting in you that we are cooking and we're working with a couple of ⁓ companies and and brands right now in the space of high consideration ⁓ kind of ⁓ verticals, like usually where y users like to ⁓ users need to make a decision that's not trivial. You it's not buying cucumber and soap, it's you know, kind of probably somewhere where, you know, you're gonna spend your holiday in or you're gonna you know, buy something that's that's non trivial pro priced and kind of all that. So so some ⁓ good retail brands and some travel brands. Brands and all that. We're building very interesting next generation AI native interfaces for that. So if that resonates with any of your ⁓ listeners, if kind of especially the ones that might have an inventory, larger inventory of products or services ⁓ or content even, I would love to hear back. Kind of, yeah, we ⁓ we are we have a small number of pilot ⁓ spaces ⁓ open and yeah kind of ⁓ together we can you know define the future of AI native interfaces. Ben: Okay. Mm-hmm. Amazing. Well, we'll get the word out. mean, I can think of a few people at the top of my head, but I mean, I genuinely think, I mean, having used your app for a little while now, ⁓ it's so much clearer and so much easier to navigate than a lot of the more traditional AI options out there. So I fully expect that this will be the way that people interact with AI going forward. So. ⁓ And I think it'll be really exciting, not only just in our space, but in kind of every space. think the kind of the UX of it will change a lot of the way that the various industry industries and consumers interact with AI. So ⁓ thank you for giving me the time and thank you for coming on and explaining everything. It's been a genuine pleasure. Milena: My absolute pleasure. Thank you. Thanks so much.