roxana-gharagozlou-from-feature-factory-to-delight-factory

[00:00:00] Hello, thank you very much for being among you today. Thank you for the invitation, Roxana.
[00:00:08] I'm Roxana Garagovnou, so I'm Head of Product Design at Bedrock Streaming, which is the platform and streaming engine of the RTL and M6 group. So, I fell into the Agile pot when I started my career. Since I joined the firm that many of you know, O'Tech Technologies. And uh, the design and product department. And uh, I've never stopped learning my craft because, ultimately, for 18 years now, it has never stopped evolving. We've transformed our organizations, uh, we've re-examined our product practices. All the web trends, all the digitalization, and now, AI.
[00:00:57] So after a few years of doing some consulting, I decided to go back to product companies. To try to work on this famous impact that we already talked about earlier, and really try to scale our practices, to transform our organizations from within. And work iteratively on what could bring value, questioning ourselves continually.
[00:01:25] So that's it, I built my career on these transformation issues within data-intensive tech environments, and that's also why I'm with you today.
[00:01:37] So what I'm going to do today is rather share a few observations with you. Uh, give you a reading, an analysis of what I think could bring uh, value when we design and produce uh, our digital products, in everything we create. Personally, uh, it also makes me question because I strongly believe in our collective intelligence. I also worked for a few years within a product and digital trying to generate collective intelligence within large groups, and it's really something I believe in: our human dimension, as human beings, to be able to adapt. To unite ourselves around meaning and to transform ourselves, because finally, I've experienced that throughout my career and it's not over yet. So let's look together at what are the different levers that allow us to create uh, this value in products that are both attractive and very profitable. I'm going to share my approach with you on how to ultimately put emotion back at the heart of the business strategy and our products.
[00:02:51] So, a finding, perhaps one of the most persistent in our business, is the confusion between the deliverable and its effect, its impact. And I say persistent because I think there's a certain acceleration in recent times with AI. In fact, the output is the deliverable, the code, the mock-up, the PRD, the stories. That's our operational commitment. And it clearly needs to be done. But the outcome, that's really the result. We'll talk about it again later, it's the measurable change in behavior among users, which then obviously generates turnover.
[00:03:39] So we don't produce to produce, we produce to have an effect. Often on the ground, as we've already seen earlier, we have a whole bunch of analytics metrics to calculate velocity, reliability in all our production activities. But rare are the teams, product for example, that really question the lasting impact, what impact it actually had on user behavior, and if it really generated value and benefits for the company. Because ultimately, all of us, we're not just here to produce, we're here to create meaning. Uh, one of the developers is not here just to generate code, produce code, a designer is not here to produce mock-ups, a PM is not here to store user stories. Together, we are here to reflect on issues, find solutions, for example, uh, with objectives that resolve around increasing uh, recurr- recurrence, rather, uh, adoption or uh, other metrics of that kind. So concretely, to make sense. It's really a topic on which. Uh, I attach a lot of importance to uniting ourselves together and creating meaning in the experience that is relevant in the products we develop.
[00:05:06] So, when we talk about this first observation, which was confusing the deliverable and the value generated, we add on top of that a second.
[00:05:15] Amalgam, which is confusing uh, quality with an accumulation of functionalities.
[00:05:24] Because a product, the quality of a product is not uh, measurable by the number of available functionalities. And from organization to organization, this amalgam constantly returns. So yes, obviously, a large part of our activity, as individual contributors, is to produce, produce, produce, produce. We can even invent a whole bunch of methodologies to produce faster. And we even get a certain satisfaction from producing on time, from having closed so many tickets uh, in the sprint, uh, on time. We measure that, it's necessary. But that's not what makes the difference.
[00:06:05] Because also, on the other hand, we have top management that puts pressure on us to go even faster. And what does that mean? It means we have a real tendency to want to do exactly the same thing, but in a shorter time frame.
[00:06:19] And precisely, when we confuse volume and relevance, uh, and we think about AI, it's not surprising to imagine that AI is a kind of Eldorado. Which will allow us to accelerate everything and therefore have an even greater product quality. Except that, do our products generate more value than functionality? I'm not sure. And so, does AI keep that promise, that Eldorado?
[00:06:48] We ask ourselves the question. And precisely, I saw a Gartner study recently that I found very interesting.
[00:06:56] Gartner estimates that more than 40% of AI-driven projects will be abandoned by 2024. Due to the escalation of costs, due to an uncertain outcome, and due to insufficient risk control. So, profitability, debt, these are questions we ask ourselves. Because we all practice AI more or less intensively and recurrently, there's a real trend there, and uh, we have a real adherence. And we all want to try it, to play with it, to test it.
[00:07:32] What really questions me is also seeing injunctions to use flourish in organizations. Use it to use it. We even have calculations uh with KPIs, there are metrics on the number of tool volume that you use per day and that each team will use per week to make sure that everyone has correctly taken the AI turn. So a form of race for self-promotion, ultimately, we still confuse volume and relevance. All of this questions us. And indeed, what it creates is also more debt when the processes are not in place. And they are not in place since we are obviously registering the normal. So, it's a debt generator. So, I'm going to let the specialists talk about technical debt, you know more about that. I'm going to talk about another debt. I'm going to talk about experience debt.
[00:08:31] I'm going to tell you about a project that I led some time ago, a study on 8 to 10 hotel booking engines.
[00:08:43] So, I had to compare several booking engines, these products really established on their market internationally.
[00:08:53] On paper, when you compare them, the functionalities were almost identical on each product. That is, uh, did we have the ability to display public prices, member prices? Did we have... Did we have uh, a whole system in place for promotions, with percentages, with different types of currency, therefore several languages, fixed prices, the ability to pay uh, when we were a member of the club? All sorts of functionalities that ultimately, on paper, were identical.
[00:09:30] But when we looked at the journeys, in fact, the data told a whole different story.
[00:09:39] So I've simplified it, I've outlined it, I'm not going to show you all the paths. But I really studied all the paths of each of these solutions because I had to create one myself, and I really wanted to have a global view of the system and of the competition. So, I'll give you the diagram, a typical diagram: I search for a destination as a user, I select a hotel. I find the room and then, uh, I compare the rates. In most cases, obviously, the member rate is prioritized, it is placed at the top. And it's often on this one that the user will click. Arriving after several steps, the extras, all that, arriving at the checkout moment, what happened was that we couldn't complete the reservation.
[00:10:31] Why? Because we had selected a member rate, because this member rate required having an existing account, or else we weren't logged in. So I'll take the case of the non-existent account. At that moment, what the user does, if he's really motivated, well, he clicks to create his account. And there, he goes to another URL.
[00:10:55] He goes through a whole bunch of steps to create his account, he even has to go back through his email, and when he comes back, obviously, the cart is lost.
[00:11:05] He has to go back to the home page, since he was on the booking engine, he does his whole journey again. In the worst case, he has lost the room, the rate has increased, disaster. So, at each of these very complex stages, in fact, there are drops in abandonment. The user leaves, it's less money, it's frustration.
[00:11:27] And that's something we really worked on so that it wouldn't happen again. Even if the features were equal on paper, the experience was catastrophic in most cases.
[00:11:37] Two actors stood out, and we tried to do even better from an experience perspective. So, what did we do? Well, we worked in a team. This is a point that comes back very strongly in the methods that I try to implement, which is really to work the tech, the product, and the design jointly. Because the number of possibilities and capacities comes from each side.
[00:12:06] You can't work as a designer in your corner and have a great experience if you don't know the technical capacity, the infrastructure. It doesn't work like that.
[00:12:14] And so, that's what we did, and we delivered something extremely powerful. Especially on the hotel side, we had over 40% conversion rate, we were delighted.
[00:12:26] So that's the transactional part. We optimized the transactional part.
[00:12:32] And suddenly, that's measured directly.
[00:12:36] Now I'm going to tell you a second anecdote that is related. It's precisely one of my colleagues who tells me that, well, we have several offices and he often comes to visit us in our offices and he always takes the hotel that is next to the office. Obviously, it's the closest, it's the most practical. This time, uh, he couldn't book, there were no more rooms available, so he decided to exceptionally go to a hotel a hundred meters further. And arriving in his room, he finds a small note, handwritten, handwritten, that's rare, handwritten, where it said Happy Birthday. And it made his day. He was delighted, he experienced an unexpected emotion and really, uh, it changed his day. And it changed more than his day, it changed his behavior.
[00:13:27] Because now, when he returns, he books at the hotel 100 meters further. Why? Because he adored that experience. He completely adopted it and now he books there. And that's what we're going to talk about now. And it's this strong and positive emotional experience that makes the user or the client change their habits. And that's what we're going to try to look for when we build a product. And I'm delighted to talk to you about product design, and I'm delighted, Christine, that you're in the room because you really inspired me a lot. So Nesefrine, a product leader at Microsoft, Google, and Spotify, uh, set up a whole bunch of studies and created what are called uh, studies on product design. So, to get the bag out of its context, uh, Nesefrine, if she's in the room, is here. She set up three types of design.
[00:14:30] First, the Lo-Like. It's a product that fulfills the functional promise we came looking for. It works, but in a saturated market, Lo-Like is not enough to make a difference.
[00:14:46] Next, we have the Surface Lo-Like. That's more emotion, but without the function. It's all the little frills, the animations you encounter in your experience on the services and apps.
[00:14:59] It makes you smile, but it doesn't make you come back. And the third one, the D-Light. Here, things really become strategic. It's when a functionality meets its functional commitment, but also creates a strong emotion.
[00:15:18] An emotional experience, and the two together. we came to look for. they share, but in a saturated market, the delight alone is not enough to make a difference.
[00:14:46] Then we have the surface delight. That's more emotion but without the function. All the little flourishes, the animations that you encounter in your experience on the app services. It makes you smile, but it doesn't make you come back. And the third, the deep delight.
[00:15:06] Here, things really become strategic. It's when a functionality responds to its functional commitment but also creates a strong emotion.
[00:15:18] an emotional experience, and both together. And the figures behind this third level, the deep delight, are quite telling. We have retention multiplied by two, revenues as well, and above all, very importantly, recommendation rates that go up to 60% more. That's, that's very beneficial for your product.
[00:15:43] And there you go, I recommend reading this book to you.
[00:15:49] It really inspired me a lot and you'll see, I hope it will inspire you too.
[00:15:55] To continue on the delight product, now, uh, how can we integrate it into our conversion strategy? that generates revenue.
[00:16:05] So, it's driven by several factors. In fact, there's a three-dimensional system that works together, a bit like a multiplication. The first is desire.
[00:16:16] It's the deep delight in action.
[00:16:19] It's what makes you want to use this service or this app, uh, and this product. Even before you started using it. So if you think of big brands, big names like Apple, Nike, Pokémon. If the desire is absent, in fact, the rest is useless. You can have the most stable platform, the best code in the world, it's useless.
[00:16:44] The second, so there 60% we talked about recommendation rates. The second is trust. This is largely generated by customer reviews, social proof, the feeling of belonging to a community. when others have already chosen this product. This lever can improve your conversion rate by 20 to 35% on average. And the third is ease. We saw it in the example with the hotel reservation engine. There, with fluid journeys, uh with quick check-outs, loading times optimized, we can improve conversion by 20 to 50% in funnels and well-designed experiences. It's really these three points that need to be put in place when creating our product. And it will be important to take it into account when building our roadmap.
[00:17:43] There are two levers in the deep delight that are very powerful, so it is, on the one hand, the friction lever and that of desire. Two fundamental levers, but of a distinct nature. Removing friction is unlocking your user. Your user had a clear intention. He came to buy sneakers on the e-commerce site, he already had it in mind, he didn't necessarily have the exact model but he had an intention.
[00:18:17] Your friction prevents it, you remove it, conversion immediately improves. That, in terms of product, should be ease and the foundation. Creating delight is really different. We saw it in my colleague's case. create a positive experience, uh, wait, I went a bit fast. Creating delight for the colleague, uh, his user experience, is precisely about seeking that emotion and that desire. And, uh, social delight is one of its very powerful components.
[00:18:53] But we can think of some examples, for example, of products for which we were able to accept significant friction because the desire supplanted the friction. I'll give you a somewhat comical example, but I don't know if you're aware of all that had to be done and undone around Pokémon and the latest collector's tracks. Uh, I've experienced that. It's the idea of waiting all night to be the first and being sure to get it. In fact, in this type of case, there is such strong adherence and desire that all the hardship of waiting all night, even in the rain, changes nothing. It's the same when we want to go see a rock star concert. There are a whole bunch of cases like this one where ultimately desire supplants friction. And it's important to remember that when we have to make arbitrations on the roadmap. What ultimately generates the most adoption, retention and business impact for our product. So remember that. And the question to ask yourself is, where do you position yourself in the products you are building? Are you more on a transactional market, uh, like e-commerce, booking? For the vast majority, users therefore know a little about what they are looking for. And your absolute priority, uh, in this type of transaction, is not to destroy their intention. Exactly, feedback like Amazon's, for example, which built its entire empire on this type of subject, with one-click checkouts. By memorizing your bank card with very easy returns, ultra-predictable deliveries with the delivery person sending you an SMS. Zero obstacle between intention and the act of purchase. Otherwise, there is also, uh, the affinity market. The affinity market, as we have seen, is truly a brand universe, a community, a cultural universe, for example. where people are not just looking to make a transaction but to experience a strong emotion. And in these cases, your main lever is really desire. Emotion can even take precedence, we saw it over efficiency.
[00:21:21] Try to see a bit where you stand in your products that you are building.
[00:21:27] I'm going to give you examples of leaders who, on the market, neglect neither. Neither the transactional aspect nor the emotional aspect. Since I work in the streaming universe, I took two examples from streaming, but from musical streaming.
[00:21:46] Spotify and Apple Music. Two products that have two very different deep delight strategies. but, uh, which also work around emotion and social connection between individuals.
[00:22:04] So first Apple Music. Have you ever listened to a song on one of your music apps while trying to see what the lyrics are?
[00:22:17] Has that happened to you? Have you even tried to sing along?
[00:22:24] Well, so those who know Apple Music probably know a little feature that has a microphone on the side. And with this microphone, what you can do is modulate the intensity of the singer's soundtrack. And if you start to feel comfortable when you sing, you can even progressively remove the singer's voice. And what happens there is that you can really enjoy what you're doing, because you really feel like you're at karaoke.
[00:22:55] What's interesting about this is that Apple has also based its experience extension strategy on its entire range of Apple products, its entire physical product ecosystem. Since with this functionality, if we also have an Apple TV, what we can do is directly cast that session to our TV and we can really, then with the other features that come to enhance the experience, for example backgrounds with stand-lights, voilà, all while you can really believe you're a star on stage. with a lot of fun, alone, but as we know, the karaoke experience is often with others. It's with your group of friends, it's with your family. And what happens there is that we have completely created an experience because we forget the object, we forget the functionality, we are really in a lived social experience with our friends. The biggest brands will even say "Hey, let's have an Apple Music night at home." There, we are truly maximizing adoption. And there, to fully understand this emotional factor in product adoption.
[00:24:10] Now, I'm going to talk about the Spotify case. Spotify has also worked on this social aspect but a little differently. Instead of working on the Apple device product ecosystem, they worked more on connections to different other apps that are part of their users' lives. For example, if you post on Instagram, then on Instagram to choose a music that goes with your story, you're going to pick from the Spotify archives. Another app and connection with the social ecosystem is the connection with the dating app. The famous dating app ensures that when you create your user profile, you can sell your top musical track there. to enhance your profile, your musical personality.
[00:25:10] And this question of musical personality, Spotify, as we see here, has worked on it with the record, uh, in fact, has worked around the wrap feature.
[00:25:22] It's in fact, every year, millions of users who share on social networks and who flood social networks with their playlist of the year, their statistics, millions of organic impressions generated for free because the experience instantly made them want to share. So here, virality and social are obvious in this.
[00:25:52] We really see two distinct strategies but which revolve around these same notions of emotion, delight and social.
[00:26:06] So yes, for my part, AI now with everything we've just said. Today, we know, AI allows us to generate in a very short time what used to take several days or even weeks, it's very powerful.
[00:26:22] But this time gain should not therefore discourage us from piling up functionalities, but rather allow us to work on this delight, this emotion that generates adoption and profitability.
[00:26:39] But actually generate all these studies that allow us to really pinpoint what will, uh, make users change completely? What will be the triggers for adoption? What will make us not churn? All that is possible if we do in-depth studies, if we cross-reference all our data. If we cross-reference, uh, the data, uh, from our CRM, the data, uh, from the developer, the data from store reviews, the data from user studies, interviews. And in that, AI can really help us. When we finally analyze all our qualitative and quantitative insights, we can really cross-reference the data and generate opportunities. But today, what happens is that often, uh, we have a whole bunch of documentation that's just sitting there, in fact, in our own repositories. Because we have trouble interpreting it. Let's be honest, in terms of labor, it's really complicated to generate that. It requires dedicated teams for that. But imagine now being able to automatically cross-reference, in a very short time, all your user research, your bug tickets, your store reviews, and analytics reports. Imagine that this analysis could even highlight patterns for you and allow you to understand that in such and such a case, we are a customer at risk and there is a high risk of churn in the next six to nine weeks. Let's imagine, uh, that we can understand what the triggers for adoption are and together, uh, really create dynamics that will lead the user not to churn and to adhere and to adopt and to talk about the product. Imagine we have all that data.
[00:28:38] With AI, it's now possible to work on these aspects, it costs much less and it's very qualitative.
[00:28:46] We can finally segment our users based on their true behavior to build these opportunities.
[00:28:55] So AI can be the driving force of this common language. We can all work together with AI to have this impact. This reflection that allows us to come together. I really consider us, with the evolution of our professions, as product people. If we really think outcome and product effect, if we reflect on the metrics and parameters that allow users or customers to adhere, ultimately we all become product people with different expertises. but we all think products. And I think AI can allow us to rally around that.
[00:29:34] For me, AI is really a facilitator that also changes our paradigms, and I think we can all unite around the meaning and this search for a new, new approach. we can all work together with AI to have an impact. This reflection which allows us to bring us together. I really consider with the evolution of our professions as product people. If we really think about outcome and product impact, if we reflect on the metrics and parameters that allow users or customers to join, ultimately we all become product people. with different expertise, but we all think product. And I think that's what allows us to come together around that. For me, AI is really a facilitator that also changes our paradigms. And I think we can all unite around the meaning and this search for new opportunities, new approaches.
[00:29:47] A few weeks ago, I posted something on LinkedIn where I wanted to talk about it because I found an incredible example, a very interesting analysis. A post from Bill Bob for Levard, Chief Citizen Officer at Nisha, who proposed a study of the Pokémon Go phenomenon. Are there any Pokémon Go fans, or have you heard about it in the room? Yes.
[00:30:13] So Pokémon Go, if you had to talk about it, what would you say it is? It's a A game. Is it a game? It's a game. Well, actually Pokémon Go is a game for those who use it, but in fact, as a product, it's not a game. It's an ultra-powerful data device. A device in which we track the behavior of users, all their movements, their in the real world, all the merchants around them.
[00:30:46] And in fact, with Pokémon Go, the business model is incredible because they created a system in which it is the merchants who are the zones of interest, the zones of interest of the users. who pay to be part of the system.
[00:31:05] So, Pokémon Go didn't win because it generated a bug-free experience. But because in fact, they created a world in which there is an emotion of hunting. In which the users by millions are inserted and in their daily time, in fact. The live has really been conceived, but they worked the whole business model around this live so that the product becomes profitable. That's what's incredible. And we can go very, very far with these live product stories. We can completely work the business model of the product. And that's also what AI allows us to do: to understand all these factors and all these triggers. We can even pivot its product, because the best products are made and remade, thanks to these insights and this understanding. That's why I'm really saying that emotion is not just a negligible factor, it's not the icing on the cake. It's really a product strategy. So how do we do that now, concretely? First, the teams must stop optimizing their user journey on their scope. Otherwise, we completely lose sight of the global experience. We have optimized experiences, but in fragments, so the experience when it's fragmented, it's a drop rate and it's a loss, a loss of convertibility. So, stop focusing on our objectives or our deliverables, to use the same metaphor as before. Stop just focusing on our deliverables, but really think global experience. global and the impact on its users and therefore on the user. And everyone must embody this, in fact. It's not just the management who have to push this. It's everyone who has to embody and understand and transform themselves to have at heart to measure this impact together. Another condition for it to work is really to set up a guarantee of the product experience in a transversal way. Because precisely, the more teams there are, the more there is a risk that each team works on its own scope and that the strategies are not completely aligned, and that damages the global experience. So, there must be a transversal governance of this global experience, both on the transactional level that we talked about and on the emotional level. And to recap, so my key takeaway, the first thing is to think about objectives to achieve, value rather than deliverables. Value comes before volume, remember that. The second thing is to quantify your experience debt, open and re-ventilate, give new life to your data. analyze your conversion funnel and really put your finger on what is a game changer in the product compared to the real experiences and the reality of your users, of your customers. And the third thing is to tell yourself that ultimately, is the product really an accumulation of functionalities? Can you really try to help your users, or even create a strong emotion? That's important. It's up to you to choose in fact, if you invest on this or on that, or if you work on both. And don't forget that AI is not a goal in itself. Because it will intensify our practices, and if we don't transform ourselves, if we don't transform ourselves, we will just fail the test. A final message to conclude is that ultimately the live, the live factory is not a project that we decide in several months. It's a decision we make every day, it's a decision when we prioritize what we're going to work on. So the live is a powerful lever. uh of conversion. I hope that I have demonstrated it well here. The best ones have therefore all well integrated and understood, and I hope that it will be useful to you tomorrow.
[00:35:31] Thank you.
[00:35:40] I think we have questions. I have a question.
[00:35:56] Thank you. I have a question that might sound a bit...
[00:36:01] uh today, on the market, we have, uh, web navigators.
[00:36:08] It's still the web navigator who has Pokémon Go. We'd like to know who is behind it.
[00:36:17] And my question is a bit complicated, but how do you see the live for the future, the future that they don't have, but they have.
[00:36:29] So, the first question is, if you think live, you think emotion, because can AI feel something? Well, I've told you, I really believe in the power of this, something that weighs heavily in all its practices. and in the way we put this ethic in place. So, really, the question of whether AI will be able to create AI. I don't know. We'll see if AI is capable of feeling something. Personally, I hope not.
[00:37:00] I hope it doesn't happen. But after working on all these questions with AI, I think it's possible. Yes, because we have AIs, we have agents who will help us do such a task, reveal such a thing. And we can have triggers that say that if this, then such a variable to put in place, so it's possible. I don't think it will think for itself, but we'll see. It's only the beginning. So we'll see in a few years, perhaps, how this will evolve.
[00:37:38] Another question?
[00:37:49] Yes. Thank you for this interesting presentation. And I'm wondering, I'm thinking about the example you gave.
[00:38:00] Uh, and in fact, you just made me think about that, and my question is, this live. Well, how can we explain it to someone who doesn't use it?
[00:38:12] In fact, if we think about it, I really showed you examples of big companies. But if in all universes, banking, insurance, it works. The live is not necessarily something that we really have to conceive and create to really show the mechanics. You see that really, uh, it's obviously a social thing. The live is obvious. When we are, for example, on billing, there are people who already have difficulties in their work, who have to enter a lot, a lot of notes, expenses, invoices, all that. The processing is very, very, very painful in fact, an interpretative penalty in a repetitive way. When I had worked, we were on agricultural credit, billing, there were two agents and it was the beginning where we took a picture of our invoice and it directly entered the lines into the system. That's what live is. Concretely, uh, you're going to gain time and an agreeableness that is really very strong. That works functionally and operationally, that works. Imagine the pleasure of having to do something like that. For someone who has done that all day, the live, yes, when I worked in Fintech too. Uh, we really had a small back office where people were piloting their profitability and their figures. We could work with the platform of self-recommendation of things to put in place for such results. According to the metrics and studies, we were told that in X% of cases, those who put this in place then generate, we have an estimate of so much additional benefit. So yes, it works. It works in our tools, in our back offices, it works. customer and for the user, and also for content, to have recommendations. of what type, for example, of visual works best, what type of layout, what type of way of communicating, first person, or plural, there you go. There are a whole lot of things that our tools can do: self-recommendation, diagnosis, and evaluation, and self-recommendation, a bit like a co-pilot in the person's activity. We can do it in professions where there is hardship and precisely that gives even more value.
[00:40:50] Thank you for the presentation. I have a question, you spoke about the transformation process, you showed us an example of Pokémon Go, but I have a question about the live that really brings value and emotion.
[00:41:10] uh how do you how do you how do you, how do you manage to do that, and how do you organize it?
[00:41:19] Yes. So, there are several models. I, I really believe in an organization that is divided with teams, with squads on the domains. on the experience funds as well. However, having transversal management is important. So having a transversal team that is really a guarantor, having ambassadors, people who are involved within the teams themselves, and also at the time of prioritization and the creation of the roadmap, it's very important. Precisely when we have a backlog of opportunities and we want to know what to focus on. The roadmap when it is unified across an entire company, all vertical components, we build our portfolio based on what we want to focus on. There is debt reduction, very rarely experience debt reduction, and that's where we need people to push it. Because we can demonstrate that reducing the experience debt is bringing profitability and that it is a very powerful lever. So there is also the acculturation of the teams on these elements. And that also goes through ambassadors, having in each team people who are at the level to be able to bring it to the discussion table. It also depends on how we work on the roadmap, because the roadmap is when we talk about it. What do we bet on? What risks do we take? What do we want to test? And what are we going to work on now, maintenance, the qualitative aspect, also reduce the technical debt, all these elements. And what percentage of the roadmap do we ultimately allocate to what is most prospective? There are a lot of theories about it, we often come to the same metrics that say that you need at least 10% of your roadmap to work on everything that is more prospective and research, to make sure that the company does not miss the bus. Because in addition, at the moment with AI, it's moving very fast. So we must not miss the boat, we must not miss the competition on the market.
[00:43:36] Hello.
[00:43:37] Hello.
[00:43:38] I'm going to ask you a question. The first one, I think it's about Pokémon. Could you briefly explain what experience debt is? And the second question is the live. You showed many examples of big boxes, of many means. Do you think that live is applicable to structures of financial functions that don't have many means?
[00:44:09] Uh, the live, uh, the first one about the question about the experience debt. Experience debt, in fact, is all the demonstration of the effect, for example, on the reservation engine. There were debts at each step. There were obstacles at each step of the journey. And that's experience debt. A dysfunctional experience is a negative experience. That's debt. If we consider that a fluid experience is the base, everything that is non-conformant, that's debt.
[00:44:44] So that was the first point. And on the live, so the question was, can we do that when we are a small structure? And there I agree, I really agree with the question, because I experienced having to abandon big research projects because we didn't have the means. The research I had done, I had three months to do it. It's huge. In some structures, you can't work like that because it represents human costs, time, money, and of course, you can't do that at home. So what do we do? We have repositories, and often they sleep, as I said. And even with AI, we can automate all these processes. We don't need to have a dedicated team for the voice of the customer that will spend weeks studying all the reports and making us, well, weeks to present us per perimeter, a report on the voice of the customer. No. We can do it continuously, we have agents who treat that, we have opportunity detection, and each time it's the right time to co-construct the roadmap, we arrive with on the discussion table. factual elements, trends, projections that make us think that if we work on it, then we will have such an impact.
[00:45:58] So yes, I think that now we have the capacity with AI to work on these issues, and that's very good.
[00:46:12] I think we're good. Thank you again for listening, and see you soon.

Transcript (Translated)