Prosus N.V. (PRX) Earnings Call Transcript & Summary

Welcome, everyone, to Lisbon. Thank you so much for joining us and for people joining in on the webcast. I have to tell you the back story of that corporate photo was actually taken by my brother-in-law yesterday. And he said to me, think of you murdering your husband and that evoked the genuine smile you actually saw up there. So the reason why this was done at such last notice is, unfortunately, Eoin couldn't join us today, and that's because he's under a couple of feet of snow. So he will be joining us tomorrow. So we look forward to welcome him then. But who is here today is the OLX management team. So one of the reasons we chose to do this event in the manner and in the location that you find yourself is because of a very exciting event they are hosting tomorrow. So we've already got the management team here at captive, and we thought we'd expose them to the financial community. We know all of the hot topics that you all are facing and puzzling over, and we hope to share their thoughts, their insights and their wisdom with you. So we've got a very exciting lineup. And as you know, it's very unusual to have access to Chief Data Officers, product and technology officers and then the CEO and CFO. So without further ado, I'm handing you over to Christian.

Thank you very much. And it took me a while to get smiling as you've seen, but I think that was probably on purpose. Welcome to Lisbon. Welcome here in the manner. Thank you very much, Kaeleen, for the introductory words. Why are we here? I have identified 3 reasons. I came in on Sunday evening. I just enjoyed the sun and the mild weather. It's probably one of those things that basically you're really looking for, and I guess you also enjoyed it. The second thing is for those who are not so familiar with Portugal, Portugal has developed into a real strong tech up. Lisbon, Porto, we ourselves have about 500 people here, mainly acting in the space of product, marketing and tech. So again, this is certainly an important reason. But the third reason is the more important. It's our CLAIM AI conference. It's a big happening. It's the first industry-wide happening in this space. There is a reason why OLX does it because we are certainly at the forefront of the development in terms of AI. We are not a listed company, so you cannot know, and that's the reason why we want to show today what we have already been doing in the past, what we have been cooking on. Tim and Andreas, obviously, are the stars of today, and we'll give you a deeper insight in what this really means. So now let me get into it. And the clicker works. Yes, the clicker works perfectly well. So we call it leading through agentic innovation. AI, agentic AI is obviously the talk of the town. It's not the reason why we're here. We don't want to be talk of the town. We want to have basically -- we are playing that role in that space for many, many years already. And I think what we want to show you is what you can do in that space and what is the speed and the pace in which you need to work if you want to be in that space. And it's moving very much away from marketplaces and platforms towards very much verticalized transactions because that's where the depth and also the benefit lies. Just to give you a short heads up on OLX. OLX is a leading business. We have been consistently overperforming. And for those who have been with us on a Capital Markets Day last year, remember very well when I was starting to talk about our verticalization effort into motors, real estate and jobs, which was a clear strategy and which also define our ambition level to grow at 20-plus revenue growth rates and also to reach EBIT margins of 50%, which Suresh, as our CFO, will continue to talk about because we believe that those are margins that you can further achieve. We have continuously built our financial momentum. We have been able to deliver 43% of EBITDA margins or adjusted EBIT margins after the first half year of our financial year '26. We have leading assets in many -- across many markets. The latest that we obviously acquired was LA CENTRALE and LA CENTRALE is also something where we are now starting to roll out our product, and Andreas will certainly talk about this. If you think about OLX, OLX is a player that is existing for very long now. And we have been constantly innovating and evolving. We have done some mistakes around OLX Autos, but we have continuously tried. This is -- we have been testing. We have been moving on the technology. And I think this is something that is super important because the market view translating into what OLX reality is, I think, is the most important thing, namely we are shaping our own proactive future. In the past, classifieds were seem to be resilient. They still are, but we believe that the vertical leaders are the data foundations of the agents. That's basically what the future today already looks like. AI is a feature. It's not a feature. It's not a feature. I think a lot of people are talking about feature because it's sort of new. But for us, agentic AI is the product. Yes. It's the product. It's not a feature. It's something that you need to be immersed into and you need to understand. Discovery Channel was the past. Now it's about partner for a transaction. So it's trying to get the customers and the consumers closer to each other. It's enabling them to basically faster and better transact, which then leads in the last one where in the past, monetization was very much seed and ad-based. Now we're getting far closer towards the transaction, which is what we call a value-based outcome-based revenue, namely closer to the transaction, closer to the lead and then also showing the customers the value that you're delivering. And this is also something that will help you not only to facilitate monetization, but probably also increase your take rate because if you make the customer more efficient, there is a reason why you would also benefit from it. OLX, as I said already a couple of times, has had a decade of dominance, especially on the -- in the AI space. We started early on in 2018, data, data gathering, data structuring, that's super important because without data structure, you can't go anywhere. The data structure back in the days already helped us in 2021 of starting personalized search. And this is something that is basically the inception, if you want, of AI. In 2024, we started with our first program, basically called OLX Magic. OLX Magic was a great product, but probably ahead of its time, ahead of its time. And we learned from that, that certain variables that were in that product was then used going forward. And we released our first AutoIQ Motors program, which is probably the only one in Europe that helps the dealer in every bit of his business. Is it pricing? Is it sourcing? Is it inventory management? Is it financing? This is the most complete set of product that you can find in the market, and we're adding new functionalities every day. On job site, in January, we released the matchmaking. The matchmaking is something that is very similar to the dealer side, but on a very different vertical. And tomorrow, we will show the world basically what our ChatGPT Compass is going to look like, namely the real estate app that we're going to use and that we're going to launch into the user hemisphere, which, again, helps us also on the auto side because in April '26, we will launch sort of the same product on the auto side. Don't forget, the good thing is that the underlying tech allows us to play around synergistically between motors, real estate and jobs, jobs, real estate and motors. So all those things are interdependent and therefore, things are working very, very well into each other and which helps us also to gain speed and to be much faster in the iteration of what we're trying to achieve. Now I've talked enough. I think Tim, who now takes over, will give you a very deep insight into how we structure this whole AI journey.

Okay. I don't know who that guy is. I'm Tim Davis. I lead up our product, technology and marketing efforts here. First of all, welcome to Lisbon. I think we arranged a special early spring day for those folks in Northern Europe, and hopefully everyone has a chance to get outside and catch a bit of sunshine. I am super excited to talk about what we're doing today. Like this is just such a dynamic time in the industry, especially in classifieds. I think we have a tremendous amount of good stuff to share and take feedback on, and we're really excited for this whole week of this AI-centric classifieds moment. So I'll talk about scaling AI-powered customer experiences, and I'll kind of lead with like AI is not a feature of the product. The core of our product is AI, but really, the core is the marketplace and how do we help customers sell things or solve human problems. I think we'll go into some details about how we think about those things. And then my colleagues, Andreas and Suresh will talk about some of the numbers that actually show the results. Okay. First of all, like eyeballs, visibility or exposure listings, beautiful business. We have made a ton of money in helping customers just find visibility or discover products. Love it. It's been super great. But that is not the future. We believe the future is taking those eyeballs and turning them into solutions or I'll say, getting to the finish line. So I think if we think about like this concept of a marketplace, actually, how do we make the marketplace more valuable, more differentiated, more immersive and actually, how do we package that in different ways. And I'll tell you what that means to us. So first of all, like we have -- I'll call it -- we have a flywheel, but I'm going to draw it a little differently today about a top funnel, which is about consumers, the middle of the marketplace mediation and, of course, the supply funnel. And we attack all of those things independently because they're all super important to us. One is like a core principle for us is people have human need. I think it's not new that for thousands of years, customers or clients or consumers have needed a place to stay, transportation, work or other services they needed to find. So one of our premises is we find customers or we meet customers wherever they're at. So there's multiple places customers will have intention. It could be mobile apps, it could be upstream search providers, it could be partner ecosystems. Our goal is to bring those all into the marketplace in our verticals and have the best place to do business. Same thing. Our goal is to help our listers or sellers sell the product. Exposure, very classic, one of the important things that we do and we continue to do. But in the end of the day, like leads lead to a sale, okay? So we're actually all about like helping customers make the sale. So we're actually transitioning to very much focused around these marketplace services are the place where the business is done or what we might call the transactional marketplace versus the exposure marketplace. And well, this is AI-driven. It's not an AI-driven, it's AI founded. So all these services that operate autonomously in the marketplace are built and delivered through AI and can be packaged in multiple experiences, which I'll talk about. Here. Okay. As I mentioned, like what we've been winning. I mean our results have been fantastic for the last couple of years on traditional classifieds business models. Suresh will show some of those numbers that we've been driving revenue and margin very aggressively for the last 3, 4 years. But then also, we actually have agentic AI in the market today. This is not a future thing. This is a today thing. So we're actually generating real money, real ROI on investments and helping consumers find better experiences today with the technology we're creating as well as sellers package and produce their products in different innovative ways. And one of the things we're really super excited about is like what differentiates classifieds and especially OLX in this world? Well, one is we see a ton of data. And I think it's what Christian mentioned is like the scale. The scale is on at least 2 dimensions. One, we operate across many geos. So we see a wide diversity of consumer behaviors, and we can take that data and aggregate it into interesting insights across many different geo markets. Second, we operate across a bunch of different verticals. I think that's one of the differentiating factors of OLX is we see consumer behavior in cars, jobs, real estate, even goods and services. So we actually have this huge pile of data unique to OLX that we can synthesize package and apply algorithmic AI to actually generate insights. That's everything from like vehicle history is cross border, something actually is kind of hard to do. But looking I actually have a lot of personal connection to is like the jobs market. We actually can see what employers actually hire for, the success rate of those applicants and actually fine-tune the matchmaking services over time based on who's applied for these jobs, how they performed, what employers come back and look for. Those are real proprietary data things that are very vertical-centric and I think can't easily be replicated outside of the OLX ecosystem. Second is like trust. Well, in any high-value marketplace experience, buying a car, looking for a job, there's the need for quality data, okay? So you trust these transactions, are you giving your personal information out to some place that will honor integrity and actually connect the appropriate services to a trust foundation and mediate those in a marketplace environment. I think about like all of the untrustworthy data that exists in the world. I think this is a key place of why actually people will come to OLX in verticals and say, well, you can't just buy fraudulent reviews from OLX. You actually have to earn those reviews, again, back to the proprietary data, back to the connection of agents and people. So again, like if we can do a search and find 5 different providers of loan services in any given market, we actually can show the history, the attach rate, the success rate and connect all those services to the transactional flow. And then on scale and depth. As I mentioned, we operate at high scale. I think we're all super clear that building this type of technology is super expensive. But what we're really excited about is we actually take this cost and reuse it and rebuild it and redeploy it across multiple geos in multiple markets. I think this is a case when scale tremendously helps us pay for this technology that smaller, more specialized vendors will have a hard time with. And I think also, Suresh, I'm going to refer to a few times to talk about the economics of this. But this is built into our business model. This isn't a new investment for us. We've been doing this for the better part of 7 years, as Christian mentioned. So our entire revenue margin profitability model is founded upon driving these technologies. We were doing AI before it was called AI. We used to call it ML, but this is the basis of how we deliver value to our customers. And I think I just kind of sum up this, what does this really mean? What's like trying to help customers get to the finish line of sell their product, whatever that may be. And we really are excited by these services being flexible, okay? We're about sellers and actually helping them sell. But we don't have a one-size-fits-all sales model. We're not a logistics marketplace. We're expecting people to put their products into our warehouse. We're saying, bring your best offer, package it, price it, choose the liquidity level that is most appropriate for your business model. You can have different offsetting ways to offer value to the customers and making that data and those offers accessible to consumers, whether they're coming from a traditional discovery experience or in the future, dispatching their own AI agents into our marketplace to find the right solution for them. Yes, and we look ahead, and we think like things will, in fact, change. Again, in a classic exposure or visibility-based marketplace, whether it's classifieds or ads, lots of techniques will become not relevant in the future, like boosted listings, probably one of the mainstays of most advertising businesses is like buying ranked listings or boosted up ad spots. When you have a consumer AI agent who's processing tens, hundreds, thousands of listings and close to real time, those things go away. Consumers will bring their own preferences, their own ranking algorithms and their own scoring to the marketplace. Our job is to make sure the agents match up in our marketplace with high integrity data and high integrity place to actually those transactions. So I think we look ahead and say, well, what will change? That's one of the things. Second is like all these services like purchase reviews, fraud bots, those are things that are intrinsically like will not be valuable as the marketplaces get more mature, generate algorithmic IP or algorithmic AI to filter sort and remove those from the marketplace aggressively with scale and deep IP. So it's a zero-sum game. There's always a cat and mouse type of thing with fraud. We think trusted marketplaces will win over time, whether they're human eyeballs or agent transactions. And the third is proprietary data. Our customers are largely small- and medium-sized enterprises. We have a bunch of data for distribution of our listings. These are things that no single job lister, no single real estate agency, no single auto dealer can really replicate. We can bring these insights and data sets to them, have the algorithm IP, which then are used to augment their listings, their value and their decision process. So we're basically again helping people sell their products in a larger context, but still allow them to personalize what's relevant for them. So I think about personalization in that context, we're personalized for the consumer, but we also personalize for the seller. And I think about like what winning looks like -- well, first of all, actually, sorry, these matter. These are in market today. So like we're actually seeing real tangible results, not just future agent transformation, but things are being packaged, customers are picking up these services, and they're showing it in our financials quarter-over-quarter. And we are relentless about doing these things in front of customer, finding product market fit and scaling up and boosting things that work. I'll just talk about one of these data points. It's one of my favorites is the jobs market or the matchmaking marketplace. I love this because we solve some really essential human needs here. One, job seekers, they may be in a bad time. They may have lost their job or looking for a better life for their family. We're helping them find better fit, more likely to find -- to pass an interview process and making that time to find that job happen that first touch point within 5 minutes. At the same time, we're solving an economic opportunity for our sellers to find more candidates better suited to them, preserve precious human time for interviewing qualified candidates and shorten the time to find labor or capital to help them make money. So it's a beautiful market and that I think is a really great touchstone piece for AI. And again, they show up in tangible numbers for our customers and how they purchase and consume these products. Another theme here I want to talk about is as we -- anyone who follows the news, the world is on fire for AI. And one of our principles is speed is the only sustainable advantage. And it used to be, we talked about being 20% faster. But now we only need to operate at computer speed or silicon speed. This is a great statistic I read a few days ago is the world market for compute will go, I think, $3 trillion of investment in the next 4 years, doubling the world's capacity for GPU or general compute. That just staggers me as a technologist. Like there's so much more capacity to do these things than ever before. And that means anyone who's operating with people first versus AI first will fall behind. We need to be operating 10, 50, 100x faster, getting value to customers, not squeezing a few percentage points out of our existing team. So again, it's a very AI-first infrastructure that we're all focused on. So again, what are we doing? Well, we're doing like more than 1,200 product updates per week. We test these, we experiment, we fine-tune them. We see what customers want. And we see something, we double down on it and scale it up, put it into a package or an offer and get it to market at scale. We love that. Same thing, the delivery rate. We are not immune to this. So again, not 20%, but 2x faster, and we expect to go even 2x faster in coming years. We have like roughly 2,600 employees at OLX. We also have roughly 2,600 autonomous agents working for us. And some of that will be 5,000, 10,000, 20,000. Our scale out will be agentic workforce as well as agentic value delivery, not with people. Well, we like to win, and we want to keep winning. So first of all, as I mentioned, like we've had a fantastic several years. I think we've actually built a culture, a drive to win and the whole company. Today is about AI and technology, but we're also about customers. And the key interface to customers and how we build technology and take it to market as well as learn is often coming from our account management and our sales teams. And of course, our marketing groups are about a bidirectional conversation. We think that's a deep moat of how do we actually help our customers co-win with us. It's not just a one-way communication. This is a dialogue, which we think is essentially a value to how OLX does business. And then kind of these categories or these rough boxes I laid them out is, first of all, we're AI first. All of our services are built with AI at the core. They might be packaged in traditional Discovery UX or into a ChatGPT experience, but they're AI core into the marketplace, building on the proprietary data and algorithmic IP that we build on. Marketplaces, we are making excellent tools for our suppliers to augment their data, sell in the way they care about. And also, we think a super interesting differentiator for us is we are verticalized. We think that the data has shown, the economic ROI has shown huge value to being a specialized marketplace and domain focused. We have people who deeply understand the real estate business, what drives agent behavior, what drives sellers, what drives apartment rentals. Same thing for motors, same thing for jobs in our other businesses. So we are able to concentrate that value, create the right products, create the right sales engagement model for our customers, which I think is very different than a broad generic approach that you see from maybe other providers. Yes. And those are things that -- it's not just technology, the whole company aligned to this winning metaphor. So on that note, I'm going to let my colleague, Andreas, talk about the details of how...

Thanks a lot, Tim, for showing us the future, how we think that the future will evolve and the trends that lead to that. And from here, I would like to also emphasize concrete AI use cases and data products that we have already in production, what comes next and also the culture of innovation that will take us there. So some of those numbers, of course, you have already seen. I would like to emphasize here that we have more than 75 customer-facing ML and AI use cases, 25 generative AI use cases and more than 10 agentic customer-facing use cases that have been developed in the last few months. Of course, everybody in this room knows that generative AI performance is improving very fast, fueled by science. On the left graph, you can see how much time it took traditionally for ML to reach human level performance. It was typically years or more. And in the last few years, how this has compacted with generative AI and more recently with agentic AI to just months or even weeks. And on the right side of this graph, you can see some of the many projections of how the agentic market size will evolve in the next decade. And of course, the numbers are staggering. Of course, putting those together, it means that we are in a transition phase. The typical static experiences that users were experiencing in the marketplace, searching, comparing items, purchasing through static interfaces is largely going to change in the next few years, where agents are acting as consultants and advisers for users, helping them search, negotiate and potentially purchase items. And this means that the user experience in the next 1 to 5 years will be very different, more different than the previous 15 years combined, have all the differences that have been brought in the previous years combined. OLX's innovative approach to address this consists of 3 different streams working in parallel. So our first stream is called optimize, and this is improving the existing user experience and of course, maximizing profits. It's basically improving our bread and butter. The next stream is called scale, and this is all about moving fast and finding product market fit for upcoming solutions and innovations. So allowing the teams to really experiment quickly, take risk in a controlled way and find what works best for the user. And finally, we have disrupt, which is our bold bets and really inventing the future. In a world where speed largely defines the winner, our execution is going from good to great. Some of those numbers, I will not repeat because Tim has already mentioned that. But what I would like to emphasize is that we are running already more than 600 AB tests per year, the majority being AI and ML related. And those numbers are actually accelerating. So you can see that on the right part of the screen, how our speed of execution is improving year-over-year in this space. And of course, 600 is already quite a lot for classified companies, but we are not happy with that. Our aim is to actually double this year again. And this is on top of quadrupling already in the previous 4 years. So putting this together, if we rank across some of the core dimensions of having a cloud scaled AI platform and AI-enabled tech teams, OLX is fully on cloud since many years. It has a unified tech stack across all the markets. It has a quite advanced generative AI platform that allows our teams to experiment quickly with all the major LLM providers from OpenAI, from Anthropic, from Google, from 11 Labs and many others as they are coming. And it has also a very large and capable AI-enabled team and total engineering capacity. What I would also like to highlight here is that OLX has agentic AI investments and use cases across the entire customer journey and user funnel. So if we think about the top of the funnel, that is all about inspiration, exploration and intent, then middle funnel that is about comparison, trust and personalization. And finally, lower funnel, which is about selection and support. Of course, in every different part of the user journey, we have different intents. So at the beginning of the customer journey, it is all about multiple touch points for the users, for the potential buyers and of course, seller value and convenience. On the middle funnel, it is all about choice and trust. And at the bottom of the funnel, it's about peace of mind. And with this, I will present just the tip of the iceberg, 5 selected agentic AI use cases that are already live and share also the results that those are producing. But as I said, there are many, many more. So first of all, with respect to seller convenience, we have already live in all of our verticals, AI-powered ad posting, where you can already today take a photo of an item and the majority of the ad is automatically generated from that. This, for example, in real estate reduced the time to post by 50% already some months ago. And at the same time, it significantly improved conversion as well. But this is only the beginning. So our next release is going to be video to ad, where in motors, you can, as a dealer go around your car, take a video and the whole ad will be automatically created from this sort of video, making it easier than ever to post. Another use case is actually our experimental partnership with ChatGPT, where this plays to the strategy of being everywhere where our customers can be. The idea here is that we have conversational AI agents integrated in the ChatGPT flow where users can explore the market, use personalized information about their preferences, of course, and find properties that are interesting for them and then, of course, convert to buyers. So you can see here how the comparison looks between the, let's say, row OpenAI experience without our integrated app and how it looks with the OLX GPT app integrated there. And you can, of course, see that it is much more intuitive. You see the images of the listings. You get a lot of interesting insights about the properties, locations of interest, appreciation of value and many other things that it is just not possible for a generic LLM to show that same type of depth and quality of information. And at the same time, I would like to emphasize that the core user experience is still remaining with us. So everything that it plays in the carousel of the items, the order of what we are presenting there, the type of insights that are provided is very much still generated by OLX. Another use case that I would like to talk about a bit more. I think Tim mentioned it also already, but going to a little bit more detail is jobs matchmaking, and this is all about choice and trust. So here, our agentic AI deep mines candidate profiles in a complex job descriptions going beyond just matchmaking of keyboards. And this is already live. It is the #1 paid employer feature. 59% of all the engagement actions are generated by that today. It has 3x faster positive actions and 63% engagement rate for the employer in the first 5 minutes. And as I said, this is just the beginning. Going to the dealers -- the car dealers case, we have AutoIQ in place. This is the dealership operating system. So this is something that allows the dealers to really manage the entire dealership automatically with AI, and that includes everything. It includes sourcing, it includes pricing. It includes managing inventory, promoting ads and many other optimizations. And this is already rolled out in Poland with 21% week-over-week retention at 100% of the Polish dealers. And those numbers are also increasing quite rapidly. We think that it will be above 30% quite soon. Then going finally to the bottom of the funnel. This is all about maximizing impact and, of course, providing peace of mind to the buyers. You can see how a traditional UX of classifieds looks like. It's actually one of the, I would say, relatively good ones in automotive. But we are really transforming it with the add-to-video capability to something that is the next generation of experience with very much signal reach and information reach ad formats. that help the buyers make the correct decisions, find what they are looking for and engage deeply, of course, with the advertisement. At the same time, from the perspective of the professional sellers, this has more than 70% acceptance ratio from motors dealers and real estate agents and with quite high willingness to pay as well. So I would like to close saying that at OLX, we are really believing that the best way to predict the future is to create it. And with that, I will pass to Suresh to tell us a little bit more about the financial impact of AI.

Good afternoon. And you heard the great strategy discussions on AI that Tim laid out, starting with Christian laying out the strategy for the company. And then hopefully, Andreas has wowed you going through the various AI use cases. And I hope to then ground you into what does it mean on us, on financials, on value creation and address some of the topical questions that's probably first and foremost in your minds. All right. But before I get going, as I walk you through how we are positioned for growth, first, I want to take a little bit of time showing you who we are. And I don't think everybody understands us because I don't think we come out that publicly. We've not shown our numbers or our financials. So a little bit of a glimpse of who we are, probably sets the stage before we get into some of the other topics. First of all, we are the hidden giant. We are by far the #1 classifieds company of Europe, probably one of the top most in any part of the world. This year, we should be crossing EUR 900 million of revenue at about 43% EBIT. And when you translate that into EBITDA, we don't capitalize. We're talking about close to 50% EBITDA. We operate across 7 markets. We have 70 million-odd ads. Those are the markets we operate in. And so when you think about it, we're largely -- we are market leaders in almost most of the platforms, most of the countries, not all of them, barring a very small few, we are the market leaders in almost all of them. That's a pretty big company, number one. Number two, we operate in markets that are growing faster than most of the Western economy. Even overlooking Ukraine, Poland, Romania are almost 2x GDP per capita growth rate CAGRs compared to the Western peers. So we operate in markets that are growing in economy where people are buying more cars, they're buying more houses, they're renting more stuff. They're creating jobs and they're growing faster. Third, when you look at our headroom on take rates, our leading verticals of, say, automotive or auto dome, our motors and real estate verticals in Poland are either half of the leading peers or even less than half of the peers, the leaders in the market. So we have a long way to go for the amount of value we deliver, the value extraction that we do is far lesser than that of what we have. And it's important because I'll tie to tide at the end to give you a sense on this, why it's important that we still have a lot of pricing headroom here. And then we continue to be the leader of our brands. Yes, you'll have a ton of questions saying, how is that going to stay back in the world of agenting AI, I'm going to address that. But hold the thought, for now, we are the leaders, both in terms of demand and in supply. We dwarf the #2 players in the markets that we operate in, barring 1 or 2. Now I'm not oblivious to what's going on in the market. I know there's a lot of questions going on. Public markets are giving a beating to our fellow peers in classifieds. So let me try to address the bear case first. What do I think about it, give my perspective on it, and let's probably have it in the Q&A, have a discussion on that. I'll take some time to walk you through this. First of all, this is a glimpse of our supply of automotive and Otodom. Automotive is our leading Polish motors vertical and the real estate verticals. Number one, these are expensive assets. These are price-sensitive, expensive assets for any user. Number two, each and every supply that you see is a heterogeneous product. No 2 used cars are the same. No 2 properties are similar. So that kind of tells you that, there is heterogeneity in the product. It's a high-value product. Now let's talk about supply. I've given you a frequency distribution where I show the ads per seller and number of sellers. So that tells you that, say, at 50 ads per seller, we have about, say, 80-odd people who have 50-odd ads per seller. Why is that important? Because we have 87% of sellers in automotive who post 3 ads or less, and they only account for 30% of the ads. That's the kind of SMB spectrum we're talking about. And this is where -- coming back to what Tim was saying, they need a lot of AI solutions. The #1 reason that a customer goes and engages our sales guys go and engage with the customers first telling them, why don't you answer your leads? Why don't you take the calls? These are the kind of stuff that these people go through. So we provide a lot of AI solutions there. And the salespeople -- so AI solutions for the SMBs and salespeople are augmented with AI solutions as well. And this is something that we certainly think this kind of fragmented supply supplemented by our sales efforts is definitely a proprietary data moat, against any kind of LLM or agent AI world we're talking about. Second, 91% of our daily active users are from direct or organic traffic sources. And I'll address what could be the risk on the next page. That gives you a sense of how strong our brands are. It also tells you that the organic stuff, the organic searches, et cetera, is about 1/4 of our numbers or less than 1/4 of our kind of daily active users. Now comes the million-dollar question that everyone is talking about saying, you know what, agentic AI is going to disintermediate your top of the funnel. What could be that impact on that? I've tried to lay out a scenario. I may not be perfect, but let me walk you through it. I certainly think that is resilience, okay? Let me walk you through this. This is a kind of a pie chart of our traffic. The earlier page was daily active users. This is our pie chart on traffic, okay? Now the direct traffic is about 76-odd percent and organic search is about 13%. You can say direct traffic, oh my God, agentic AI is going to be the direct traffic. I would say yes. But why would anyone use agentic AI outside our product? Tim and Andrea walked you through it. For us, agentic AI is the product. It's embedded in it. You saw that kind of supply that's going to be with us and that we'll continue to augment it with all the kind of AI solutions and solutions that make them win. Why would any user go in a direct world, go and try to do it elsewhere when they already have that app that is very superior to going through that Pepsi challenge. Then comes the question, what's at risk? So let's take an example. What I've tried to do here is, today, AI search is say, 0.03% of our total traffic. And then you -- I think all of you track the stats. It's less than 2%, 3% at most for many of the other peers. Now let's take a situation that, that jumps to a big number. It hasn't happened. Say it happens. Number two, one of the agentic or the LLMs ends up being the numero uno, takes over the world. I don't know if it's OpenAI, it's Gemini, whoever else. And number three, they choose to monetize at scale. Now if all of that were to happen, I'm taking a situation saying, let's say, half of our organic search becomes paid due to an LLM coming in charging us for that. And I'm assuming a 50% increase in price because I'll have to buy a lot more traffic. Even if that were the case, we are talking about a 3 to 4 points EBIT margin compression for us. And that's manageable, and I'll walk you through why. First of all, this is indicative. I don't want people to say, "Oh, that number is right or wrong."" We looked at publicly available information. We looked and said, okay, how much of this -- their revenue is from nonclassifieds. And we are certainly in the low end of the spectrum. So we don't have that much of eyeball-led revenue. Most of our revenue is classified revenue that the agents or the dealers are happy to pay because they're getting solutions. On top of that, if you look at -- and this is where -- if you remember, when I talked about less than 5% margin compression is a worst-case risk, why do I feel comfortable? Our margin expansion in the last 3 years is pretty impressive, 21% to 43%. People would tend to think, "Oh my God, you've slashed investments." You've taken it out of the pocket, you're just squeezing the lemon, and that's what it is. That's not what drove it. Those were the times where we were still focusing a lot on paint ship, a lot of transactional businesses, doing a lot of subsidies, chasing the inner goods category. We've changed our focus. Number two, so our core revenue, as you can see, motors, real estate and jobs has gone from 65% to 70% and grown too. That's causing a lot of the expansion. Plus a lot of the tech investments that we did on replatforming, et cetera, got over there. That sets the stage for Andreas and Tim to light it on fire with AI that we've been doing. So this expansion has come without kind of -- yes, we've been prudent on costs, but it's not like squeezing the limit. That's the expansion that has happened in us. And when you look at our peer set, again, these numbers are from public information. They're not from the same time zone, some are fourth quarter, some are half year, et cetera, but it gives you a sense on what kind of relative zone of EBITDA they are. And that's where we stand. Compared to pure-play verticals, we still have a long way to go. And our journey has been without squeezing the lemon. And oh, by the way, we have been never letting down on our AI investments. If anything, we've only increased it. We have 150-plus people. We have more than 80-odd teams that work on AI, probably the highest in the -- and Andreas has talked about it. We've been investing between OpEx and CapEx plus $30 million this year. It's only been growing. So that margin has come on top of this. So that gives me comfort to say, yes, as long as our product is superior, the direct traffic stays with us, which I don't believe why they would move. And even if there is disruption in the paid traffic or in the organic traffic, we can manage it. With that, hopefully, I don't think I can ever assuage the bad case, but at least I'm addressing some of the questions you may have. Let me go to an exciting bull case that probably we all tend to overlook. AI is already a monetization or a pricing driver. We've already embedded into our product. Tim talked about it. That is a screenshot from our current package in automotive. We already have ad to video there. And by the way, it generates 8% to 10% of incremental leads. So customers are saying, "Oh, I love that. " Give me that because I get more leads. And that's causing a lot of our price increase. Now it's very tough to say, okay, this came from AI, this came from this because they matched. We're not going to have an AI product package and a non-AI product package. It's part of the package. And on top of that, we're also deploying it now in Lab and 12, where all of these tools are coming in. And that's one of the power of the synergies we're going to have with them. And it's also driving productivity at scale. And we are just scratching the surface. Right now, the margin enhancement we're getting is about 2-ish percent, but that is only the start. We already see customer service efforts are going down -- efforts are going down, productivity is going up in a massive scale. We already are doing a lot of projects so that by '28, we get that. 98% of our content moderation happens without human touch. We're getting a lot of savings in campaign management, et cetera. And these are prospects. You can get it off any of the third-party solutions, but we are leveraging all of that. And as AI becomes more and more embedded into the enterprise, we're going to see a lot more of productivity that even I can't [indiscernible]. Finally, look, there's always a question on our industry, where is the terminal value? What's going to happen to the outer years? I feel there's a lot of opportunity on OLX. We've got exciting markets and geographies. I talked to you about that. We have a lot of take rate headroom with all the AI solutions that helps us because it's already a pricing booster and AI cost productivity is just getting started, and the risks are quite manageable. So with that, hopefully, I have laid my case, and we can discuss it more in Q&A. With that, I'll hand it to Christian.

Thank you very much. Yes, why are we excited? Obviously, again, tomorrow and the day after, you will see senior product people. You will see a lot of very innovative AI people around us that will give us an insight into what innovation looks like. On the second day, we will try to operationalize the whole thing. So it's going to be people talking about what does it mean to talent. Suresh already made cases around the financial piece, but we're going to leave that there's in more. So that's the reason why we're excited. The second thing is we're absolutely excited and I'm absolutely proud of being able to run this business because as you've seen, not only are we leading, which is good on financials, but we're also leading in terms of product. We're also leading in terms of the product and also in terms of the ideas. And if I have to wrap up, it's to say, and this is, I think, very important, the proactive future, the OLX reality. We're not reacting to AI. We are leading it. And it means that basically, we are working on very deep vertical integration, which we started and real estate motors and jobs is a testimony to it. We have a proprietary data loop, which is super important because it's not only the underlying data, as Tim was saying, but it's also the structure of the data and the usability of the data that makes it so important. It's an AI-first infrastructure. And you've seen on one of the first slide, we have invested across those years, about EUR 200 million already. So we are not just starting. We have been there, and we have spent it and we continue to spend as the CFO was telling, which is always a good sign. And we have a desire to win culture. We have 2,600 people in our company. We have 2,600 agents joining us. We have many more agents that's going to join us. Obviously, those agents will have even more desire to win. And I think this is super important because this will enhance us, it will create more efficiency and will help us to deliver on our long-term ambition, which is very nicely summarized for you and 20% revenue growth and 50% adjusted EBIT margin that even if we look into the beer case, is something that we can -- I don't call it easily, but we can well achieve. With this being said, I'm very much excited to welcome you tomorrow to our CLAIM AI conference, which I think is an outstanding conference and we are representing here classifieds because classifieds an industry has not been good at that so far. And again, as a leader in that space, we are taking the helm, and we very much welcome you there tomorrow.

[Operator Instructions] Will, go ahead.

It's Will Packer from BNP Paribas. Three questions, please. Okay. First question. Could you expand a little bit on what you classify as proprietary data? In the recent sell-off of classifieds and other media Internet names, there's been quite a bit of debate around what is genuinely proprietary and what's not. In today's presentation, you gave a very fair example, which was unique inventory from the long tail of small dealers and private sellers. But could you expand a little bit beyond that as to what you would count as proprietary data to help us think through data as a moat?

I think Tim probably was the one who basically was talking very much about it. It's not only about the supply as such and basically the very small pieces. It's also the geo. So we have a lot of element that plays into it. But Tim, I think.

Yes, that's a great question. So part of it is just supply. The supply is intrinsically proprietary, although we don't always own it. But it's all of the interactions and the connections of those data sets. I think one of the examples I love to talk about is actually how are the candidates performing at a certain employer. So that's the history, that's the retention rate, the interview pass rate, even coming back and advertising for additional job requirements. We actually are building closed loops, as I call them, for employers to give us more data to capture is what's actually working for them, what are the candidates who were successful. Eventually, we'd like to capture their performance and how they're impacting the business and build a better candidate database for passive and active recruiting. Similarly, for things like real estate, it's not just like the number of bedrooms or bathrooms in the property. It's like what do consumers value? What is the comparative data, what's the dwell time, what is the actual interest in a particular distance to a park. Those are all interactions that are captured uniquely on our platform. And we, of course, database them and build insights out of those. Those are things that any one given seller can't capture. They don't really show up at the top of the funnel, either they show up inside the marketplace, what's the attach rate of a particular financing offer, what was interesting to a consumer and what actually performs better. And remember, we're a marketplace business, small differences of conversion rate are huge economic drivers. Same thing on other verticals like cars. We actually see a lot of the same cars year-over-year. People come back, resell, repurchase, relook at those same cars. We actually can track those both in market as well as across markets. We can track a car history as it goes from France to Poland, for example, what's the history, what's the accident record, all the things that don't show up in public databases nor will they show up in that, call it, a naive broad-based top of funnel search.

Secondly, on the basis of today's presentation, you could definitely envisage a bit of a change in the classified industry and that potentially consolidation will make more sense. Historically, there's been relatively few revenue or cost synergies, but in the context of some of the developments you've talked to today, they could be applied across the portfolio. You've recently completed the La Central deal. One obvious challenge to me would be the data architecture of the businesses you would acquire. I imagine that's been a work in progress. Interested to hear how the work has been on La Central from a data architecture perspective and how the synergies have played out so far?

Yes. I think this is a great question for Andreas to me at the microphone of what AI is going to transform that whole problem.

Exactly. So in the case of La Central, there is already an advantage that both businesses are fully on AWS cloud. So it is relatively easy to connect the data platforms. This is something that is already ongoing. So already in terms of financial and key user metrics, the integration basically has happened since last month. So we are still putting some final details here and there, but already at the level of financials and user metrics, this is already happening. What comes next is synergies on AI. So the first use case is that we are going to actually share with LA CENTRA are ad to video and video to ad. Those are both coming in April, and this is just the beginning. So already, we are building on top of businesses that are fully on cloud. The first level of integration was literally done in 2 months and the next level with first synergies on AI in the next 2 months.

And if I may add also, Will, the way we look into companies that we buy, it's not about the size, and it's less about cost synergies. It's what can drive that revenue growth. What is it that we basically can contribute to the top line. And this is what basically also is the fundament of due diligence process. It is the understanding of us, of our opportunity to move the things forward. And that's exactly what Andreas was referring to. It's something we identified early on. We know where L Montreal needs support, namely on the small and medium-sized dealers, hence, the video to ad and the ad to video, which is something that will help us to basically acquire them. So we're very, very thoughtful about what we're doing once we move into an acquisition.

And if I can add one more thing, I think it's thematic throughout our entire strategy of AI first is lots of integrations or things like data mergers are total disasters. They have been historically tons of people, tons of costs, unrealized value. What is different about AI and especially how we're putting AI first is we're actually making that a machine problem. We're not hiring hundreds of consultants that take 5 years and actually have a low friction. Remember, our principle here is time to market or velocity, okay? So by actually building self-described data sets, applying agentic transformations, integration, whether it's data sets across our properties or whether it's third-party ecosystem providers, we're removing slowness. There's often people in making these AI to AI or agent-to-agent interactions for data integration, API integration, things like credit scoring, things like real estate transactional support for digitization of property records. So effectively, AI is letting us bring these things faster, better to market and getting rid of the entire, a, high cost of humans; and b, slow realization of value. So I think that's an important part of how we're approaching these problems.

And then last one on all. Thanks for all the color on the Bull case. It was very interesting. I suppose one challenge for your peers is that they are listed and have financials they report every quarter, and the market is very sensitive to them. In contrast, you within a larger process, there's obviously the Tencent stake, which drives the primary driver of earnings. From your perspective, what happens to numbers during that transition where you're changing the revenue model to more outcome-based when you're grappling with the changes in how customer acquisition works, how product development works. Is there naturally a period where numbers come under some pressure before accelerating? Or is it relatively benign?

Look, we've been in this kind of AI/ML investment journey for almost a decade. So we've seen these things start panning out. So when you see an ad to video in an automotive page, it's been there for some time or something of that derivative has been there for some time. So we've seen the impact of this in the financials for some time. And the good thing, I guess, is being part of a Prosus eco family is we don't have to report quarterly at the same kind of and have to deal with volatility of external pressures beyond our control. So we deal with that with laying out a kind of a short-term financial plan and working through that. So it's that way that we do. And in terms of revenue and cost that you talked about, you could see that. I mean we do have a lot of headroom. We don't see that going away anytime soon. And the cost around traffic disintermediation, we haven't -- we haven't started seeing it. You can -- and not as any other classifieds. But in spite of that, they are facing short-term volatility.

Andrew.

It's Andrew from Barclays. First one is back to this kind of, I guess, flow in an Argentic AI world. So in this view of future where I tell my personal AI assistant to go find me a car or a house. Just curious as to why you're so confident that it will definitely be the #1 classified that is then providing that agentic flow. And is it your view that technology, i.e., your set is set up for agentic and user experience is enough? And then what are kind of the trade-offs in how much of your secret sauce you kind of open up into that flow versus bringing people back into your own ecosystem and also how the kind of commercial relationship works with OpenAI or Google or Apple or however it's going to be to make sure it's definitely you. And I guess that conversation may be different if we're thinking about, I don't know, Poland where you're super strong versus in France where LA CENTRALE is competing with Leboncoin and it's a 2-player market and maybe a bit more of a race to the bottom debate. Like how do we think about that?

I think -- I mean, obviously, Andreas as well as Tim, I think, have spoken to it, and I think it's good for them to answer. I think just one comment on the La Central. I'm not sure whether the competition with Leboncoin is exactly the place to go because that's not exactly what we're looking at. The question is always when you go deep into the verticals, the horizontal has a problem. and that's exactly what we're doing. So with the product that Andreas was referring to, the way we approach it is something very different. And that's the reason why I would probably take this into the equation, but take it out of the equation of competition. That's not the way we look at it. And that was, by the way, never look, we looked at it when we bought it. So it's something very different. But Tim and Andreas, please.

I can perhaps start with perhaps 2 points, and of course, Tim feel free to add to that. So the first point is, as I said in the strategy, and I think Tim referred to that as well, we will be everywhere that our customers are. So at the very beginning of the funnel, we want that people have choice. And if they start their journey with OpenAI or with Google or with any other place or they come natively to us, as Suresh was saying with the data points, we'll be present in all of those. I think that with the experience that we show the side-by-side comparison with the OpenAI row and OpenAI with our app integrated, we can already see that we can provide much more insights and a better experience for the user objectively. They can really find what they are looking for, see images, understand exactly how it works and of course, still see deep links that send them back to our sites to continue the middle funnel and bottom funnel experience there. But the other thing is also perhaps other players have similar strategies. But as also team and me referred to that as well, we are entering an era that more than ever, the sustainable advantage is execution speed. And this is something that we are really believing and doubling down on. We already are in a quite good place, but we are going to be in an even better place in the next 6, 12, 18 months because we are really doubling down on our existing investments and making certain that we are making the most out of what we already have, getting more and more speed as we are moving forward.

Yes. I think I wouldn't have anything to add on the technical side. But I think about the essence of your question is these agentic transactions or agentic AI transactions. And again, I think about like AI to human, human to AI, AI, we don't really care. It's going to be AI at the core and how it's presented may vary based on what users need. But it is effectively the place these transactions are going to happen. There's going to be more data to operate with. There's going to be more trust. In the same way today, you've seen things like ad broker networks actually mediate real-time ad purchasing and personalization at high scale. We look at this as high scale for our verticals. Now that's where these things are trusted. There's transaction orchestration, there's attach rates. This will be the most efficient place to bring your agentic AI and operate within a real estate or a motors or jobs or even general goods marketplace. So I think that has to happen somewhere, and we think that is the best place in a very vertically centric transactional environment with all the attributes we talked about, agentic AI, agentic data, trust, verification, history. So those are the things I think will differentiate why classifieds, especially OLX has a right to win there.

And potentially, I'm also trying to simplify, Andreas. I would say we don't give you personal assistance any opportunity to go somewhere else into us. And whether it's Poland or France or somewhere else, I mean it's not so much whether you're #1 or not. It's the way you basically are represented. And I think that's what Tim and Andreas were displaying.

If I can just supplement a question to that, just one of the ones from the chat. So Sylvia sent through a question and said, talking about your AI integrated product, what do the economics of an offering like that look like? How much does that lead cost you the economics less attractive to you versus someone coming to you directly?

Look, the cost of AI investments I laid it out, we're investing about $30-odd million on AI this year, both OpEx and CapEx. We are not trying to branch out every single investment and look at it and say, this is the return that it gives. We do investments based on priorities that we lay out and we go against it. And then when you look at the returns that we get, it's embedded into our pricing, it's embedded into our customer journey. It's embedded into the pricing that we get and the revenue growth that we see. And that's how we measure it. So right now, there is -- I don't look at it at least to say, I put x amount on this and here is the return I need to see. All I'm saying is I need to invest because this is the customer journey, this is the customer product that we need to build, and we need to be best-in-class to ensure that we get every possible traffic source coming to us. And what does that investment do? And can I balance that within the margin expectations that I have?

And I think to the point of Suresh, I think that detail is not relevant to us because we are focusing on the revenue growth. We're looking to getting 20% plus. If we're getting there, the costs that are associated with it will be born like they have been in the past. That's the first. The second thing, as I was saying, there's a lot of synergies on what we're developing between the verticals. Real estate, motors jobs, the other way around. So I think trying to be very nitty-gritty on this is going to -- it's not helping us because it would rather be focused on what is really relevant.

And most importantly, the investments that go is against the right priorities and the speed of deployment is all we measure. And then it plays itself out. Otherwise, you'll get into false precisions trying to do all kind of allocation engine on cost, which honestly, nobody can then understand.

Sorry, it was a long first one, I promise a shorter second one. So it's about data. And as there starts to be engagement in an LLM environment to take the app in ChatGPT as an example, whose data is that. So presumably, you get to keep all of the insights about how people are searching for a house or a car in ChatGPT. But how is OpenAI able to use that data that's going to get created?

So exactly, I tried to show that a bit on the example, but the way that it works today at least is the experience, the core experience is still with us. So everything that is shown on the carsell with the photos in what order you see the houses and so on. Similarly, next month that we will launch motors, all of that is with us. And that, I think, is where part of the secret sauce is. Some of the secret sauce is even deeper. So it is, as Tim said, the user journeys or the history of the cars, how many accidents they had, the appreciation potential in different neighborhoods in real estate, how that has evolved historically, but also how you are predicting that it will evolve in the future. These are quite detailed domain-specific problems at least in the midterm, it is very difficult for a generic LLM to address. Of course, in technology, it's very difficult to predict what will happen in 7, 8, 10 years. But at least the current situation and the reality of where technology is and combined with the advantage in terms of structured data that we have built over a long time, makes us quite optimistic that we will have a very strong position in that.

It's Luke Holbrook here from Morgan Stanley. One of the benefits that you think OLX has is being multi-vertical, multi-geography. And we've seen in European classifieds in recent years, actually single vertical, single countries have basically been the norm. Can you just articulate from a data point of view, what is the exact advantage that you get? Because if I'm searching for a job versus searching for a house, what is it at the back end that you think actually leads to a meaningful advantage?

Yes. Let me make my first point. I think Christian might have something to say about the overall investment case around that. First of all, a principle that I have used for 20-some-odd years, and I'm kind of an older person, I suppose, is that people are much the same everywhere. So we actually can reuse learnings, user behaviors, preferences at higher and higher scale across geographies. That's unique. So this gives us more data. And we believe many, if not most, of the consumer preferences or seller scenarios that are applicable in Poland are also relevant to France. So part of it is just the basic premise. I think that's been pretty well borne out by many other e-commerce consumer-centric companies. So that's a fundamental of why we believe this transcends any particular geo as well as any particular vertical. Again, we actually see more variety of touch points for these significant life events, searching for a job, searching for a car, selling your car. So I think we actually can compose back to what I call that 2 dimensions of both geo and vertical into a single set of algorithmic AI plus proprietary data sets that let us apply to new problems. That's kind of, I think, our fundamental hypothesis. Now as far as why one European-wide company, Christian, maybe you can talk to that.

No, I think -- I mean, just to the example, if you've seen the Pepsi example, we're starting in Portugal. We're not selling Poland. That gives you a clear understanding on how we're playing around the synergies. It's not -- Poland is certainly the largest countries, obviously. But we're looking for a lot of synergies from the other. So -- and as I always say, 80% to 85%, it's everywhere the same business because buying a house and selling a house Unfortunately, whether you do it in Germany, in France, Portugal, it's not so much different. People are looking everywhere for the same. What is the valuation? Is it fitting? Is this? What about the supermarkets and so on and so forth. And the cars is not very different. The only difference potentially between cars and real estate, one moves border, the other does not. What -- if that then has an impact on business, that's a different story, but that's probably the largest difference. And that's the reason why I think the synergistic element below makes a lot of sense. So something that we have not touched and we have not basically worked at all in terms of the AI space is you can also read from an intent. So if somebody is buying an SUV, maybe he has just got a second kit. I mean, again, we are not there yet. And I know this has been talked about in my former days when I was at Scout24. But there's a lot of elements that we have not identified yet that can help us to continue to drive those businesses. we will learn.

And if I could just put one more thing on top of this, and maybe it's a bit controversial, but like execution matters. We're talking about some fundamentals. I actually think we have a track record of executing with single product, converge data, converge operational tools across our markets. I think a lot of people had some good ideas, but not everyone was able to pull it off. So I think we have some track record that Suresh mentioned, we got ahead of this. We certainly had our own learnings along the way. I can't say everything was perfect. I think we've actually crossed -- use an old metaphor, crossed the chasm around this next now, we're seeing accelerating returns -- we're actually operating single product, multiple markets and sharing these experiences across verticals.

My second question is just actually a bit on what you described as the transition of the monetization model away from boosted listings, paid visibility, I think gaming rankings. And I'm just trying to work out how you think about that gap in the way that the transition happens and what you're doing internally to help manage what you think could be quite disruptive to the existing model?

Yes. I'm going to address and maybe let Suresh talk about this. So first of all, it's additive for us. We're making a ton of money on things like value-added services, listings, rankings, letting people compete in a competitive marketplace for placement. We love that. It's funny, and we actually -- people value it. So we're happy to provide it. These newer services that we have some fantastic things coming, for example, that are more marketing tools for the real estate business we'll be launching in the next year or so, which actually is now new revenue streams on top of -- above and beyond just listings. So we're actually looking at these as additive. And then over time, the ratios may change as customers choose what they prefer. But we're not trying to diminish anything. We're trying to say what net new revenue products and monetization models can we add to the mix and then look at that as a blended growth future. Suresh, maybe you can explain it better.

I think exactly. What Tim said, we're in a state of transition where we're not deprecating what we have right now, but instead peppering it with a lot more of AI tools and other stuff that becomes part of the package and also gives us a lot more of -- the customers get a lot more of leads and better returns from that without having to get boosted ads, et cetera, and trying to see how we can push the price and find the right balance of that. And then over time, these things -- the earlier visibility products that people were paying for will deprecate automatically as these things take over. And that's the transition phase we're going through. We are learning as we're going. And end of the day, we would want everybody to put all their inventory if possible and get the best returns as possible and not have to worry about, am I getting this or am I getting that? And that transition AI helps us a lot. and some of the tools you saw, these things are really, really easy to do for the SMB population that we talked about. And on top of that, it gives a phenomenal returns.

Any more questions?

Christian, I just wanted to follow up on one of the answers you gave to a previous question. You said to win the floor, you will not give the AI a choice, but to pick you. And I just wanted to understand what that means and what advantage OLX enjoys so that no one else can replicate it.

So I mean I will give you -- I will start, and I will hand over to Andreas because he answered the question already. But so I think the way I look at it, it's about the execution, and it's about basically making sure that the consumer because he was talking about the consumer has always access to us. So we are -- we should be present everywhere. And whether it's, as Tim was saying, AI to AI, AI to human, human to human, whatever, we need to be present everywhere. And then it boils very much down to the execution. And that's basically my -- that was my analogy or my metaphor when I was saying, I'm not going to leave them the choice. And today, also don't leave the choice. Otherwise, I would have not been -- we would have not been able to build a brand like OTOMOTO, Otodom, Storia, AutoIt, et cetera. So if you want the mindset should stay the same. Obviously, we need now to make sure that the technology that is underlying is changing and that our execution goes against it. But that's the idea behind.

Exactly. Just building on what Christian just said, it's a couple of different things. So one is offering the best possible customer experience in every different channel. So if you are in OpenAI or if you are in Google or in any other, you basically have to have the best possible experience for the customers in each of those. So that's one thing. The other part is being actually a first mover in each of those different channels gives also a first mover advantage. So that the young generation that gets to interact with you, interact with already strong brands, gets the trust and then you are basically very difficult to displace in addition to that. And we are moving very fast. We are integrating with all of those different channels with all of our verticals. We announced in real estate, motors will come practically next month, and then we are moving to jobs. So everything will be there very, very soon. And we believe between that and between having very good user experience and strong brands, we will create the habit for the users that they come naturally to OLX to continue the user journey.

And probably also, I mean, Tim, maybe you want to comment, but we have also tons of content which depending on how you basically package that content, again, this will help us in different channels to steer people towards our brands.

Yes. I think that's a good point and something we really didn't talk about today, but with Lawson Trial, we actually have like a very large amount of consumer content, which is adjacent to the core listings business. And we look at extending that, packaging it. And that type of content is essential to things like the interface to upstream chat LLM-based systems. So we're actually looking across a broad things of not just listings, but the augmented data, how do we offer that up or present it to different consumer experiences make a broader funnel. Actually, I don't want to say broader funnel. We think there's going to be dozens and dozens of funnel. That's actually a core reset for us. That's a bet. It's a strategic bet for us. But we don't believe it's going to be one single ChatGPT funnel in the future. We think that the market will be healthy, vibrant and competitive and being able to package different experiences, different customer intents, engage people wherever they are in that buying journey. It could be long lead automotive enthusiasts. It could be job training and development for folks with exposure to local market trends. We think that's an essential part of our overall consumer experience strategy.

Sorry, can I just ask the -- I think the point is I can understand from a human user experience why somebody would choose you, but why would Andrew's avatar or agent choose you because the brand wouldn't matter.

Okay. Can I start? So the difference between Andrew is who is a human and has emotions and the machine is the emotion. The moment you take the emotion away, it will be even easier because the machine will always choose the best solution. The human will always potentially define a best solution as something he likes. That's the reason why I'm less concerned about the machine. But it's -- sorry for that.

Yes. No, no.

Thank you. I think that's very helpful.

Yes, I think it's a good point. I think if I just sort of kind of say that's absolutely true. And that's why I would call like the agents will be purely rational, but they will also need to operate off of data, okay? So we actually can package and augment and give a wider set of package data and services to the agents coming to our marketplace. So even if it's like headless or Zero click or Xero UX, those agents need to operate against something interesting to differentiate and choose from. So we think the solutions of more selection, more services added, deeper data stores actually will actually be the reasons agents come to our marketplace and make those choices, even if it's purely logic, not emotion.

Exactly. So just building on top of that, at the end of the day, the fundamental aspects are still choice, trust and in some sense, pricing. And we intend to have the best possible combination for those. So exactly as I think both Christian and Tim mentioned, if you have the best choice on those, it's actually even easier for the agents to be attracted to your marketplace. And I think we are quite clear that we intend for OX to be the place where all of that is happening. I think it was one of the opening slides from Tim's presentation today.

It's Giles Stone from Jefferies. Is the importance in brand investment going to go down?

So that's a great question. No. Okay. Now the packaging and the connection to the value chain may change. The relative spend that we make, we actually vary even today market by market. But when you talk about consumers and actually discovery and trust, brand, I think, is an important part of that. And the brand is not just like a marketing brands like the entire promise of the product and the realization of that product. I think those are the brand attributes that generate, I'll call it, a virtuous growing brand. So we think those are important. We think trust is a brand attribute. And those things I don't consider.