FEBRABAN TECH 2026 | PAINEL - Da ideia à escala: o que as fintechs vencedoras estão construindo

Ladies and gentlemen, resuming the Arena Fintec program, this time we'll have a chat with the winners of the Arena Fintec 2026 competition. So, to talk about what we achieved, their evolution over the years, and also about some commonalities between these two companies, I'd like to invite Luiz Lobo, CEO and founder of Falk, to the stage. A round of applause for Luiz. Very well, my dear, make yourself at home. And welcome to Fernando Dipinho Gomes, CPMO and partner at Car Big Data. Let's go. Awesome, huh? They won one day and the next day we're already here doing the panel. So, it's great to be able to count on you guys. Uh, I wanted to ask you guys, are you listening to me or not? Ah, it's with this microphone, look. Let's go. So, first I wanted to hear a little bit about you. You gave a 5- minute pitch, but I'd like to hear in a minute and a half what you do for those who do n't know you yet. Good afternoon everyone. Can you hear me okay? Yes, excellent. Just to confirm. Ah , quick pitch. Car Big Data, we maintain a network of devices that transmit images of vehicles driving on the street. We process 1 billion frames per day and form a Big Data of vehicles with a behavioral bias: location, time, region and date of vehicle circulation. We have a 12- month history and we also pay close attention to the context of the image, whether the vehicle is associated with any commercial use, whether it is broken down or what condition it is in . We created behavioral data that is unique and exclusive, because the main source is the cameras, the devices. This network generates unique data that we apply in cases of recovery, fraud prevention, claims, appraisal and location triangulation in the formalization of vehicle financing. So, everything related to vehicles involves us, observing this data through images. So people understand, those cameras aren't your property, are they ? It's a mix, few are, because we don't want to be an asset-intensive company; We prefer a partnership model where we remunerate those partners. Ah, I see. Through the monetization of that image data. Brilliant. Okay, Globo, tell us. Brilliant. It's a pleasure to be here. Well , we were born, FTO was born from the idea of ​​being able to help people solve their problems, because we as consumers, right?, in most companies we don't get very good service, right? And our mentality was this: " Wow, I come from Itaú, from Stone, right ?" So I come from that world, I'm passionate about customer service, about the customer. And the idea was, wow, how do I get people to receive better service at all the big companies? We can try to help people, but we can use the companies themselves to help, can't we? By offering a very good service, these companies achieve very strong operational efficiency and, incidentally, a very high NPS. And so we did, we created a powerful platform, initially for payments, right? We closed, well, a big project with Google and several football clubs, but everything went wrong, right? That's the life of an entrepreneur, isn't it? And we end with large corporations. Right after the pandemic, with a very large demand for the conversational payments solution we had, we then created a customer service solution. Naturally, we migrated to collections , where we developed and brought in very large brands. And we take pride in serving many millions of customers per day, with a good NPS. We don't have any customers with an NPS lower than 70. So, that's a really high NPS, isn't it? And this is where we managed to incorporate AI. And it's not this AI of today, we were already selling AI a long time ago, years ago, before Bing, before GPT and all that. But AI can be very welcoming, very human, right? So our focus is on that human connection, right? Treat the customer gently, right? With respect. And with that we have grown a lot, because we provide enormous savings to the client and great respect for the people who use it. And, Lobo, since I'm here with you, tell us about that change of strategy. That's when you turned your attention to collections. That wasn't the proposal before, only for Yes, our platform was a platform, actually, this is very interesting because the world is changing quite a lot now. We created an incredible banking platform with beastly security. to make payments, right? It's omnichannel, it connects to IVR systems, any phone network, WhatsApp, we're partners with Meta, and so on. This platform is very powerful; it understands anything— images, text, you know? It had a huge competitive advantage. Then ChatGPT came along and gave us a boost. Uh-huh . And then the cloud began to arrive. So, many people today have the ability to develop some of these things, but when it comes to collecting payment, you face a much greater complexity, right? So, what we're seeing more and more is that you used to have a much more SaaS model, right? Software as a Service, and now the Service as a Software model, you know? Americans love those phrases, but the fact is that the customer increasingly wants the finished solution. So, it's like that story, right? I don't care whose egg it is, just give me the egg. Uh-huh. So, our point now is that, and what we noticed is this: "Hey, wait, our billing system is amazing, it's better than what's out there on the market, isn't it?" And we recently discovered that we are leading this market . Last year we processed 1.4 billion in payments through us , is n't that something? 600 million in '24, 1.4 in '25 , right? And this year, undoubtedly even more so . So, we noticed that we did a very good job, a demand that the market consumed, right? And now, unfortunately, the country is not at its best in terms of delinquency, which generates a colossal demand for us, because the solution offers a very good cost and spectacular collection efficiency, right? That's why we're with all the big banks, and in the meantime, you have a good NPS to keep the customer in-house. Mhm. So, this migration was more of a market that wanted the complete solution, right? They say: "Friend, it's very interesting because I have banks that are asking for solutions where I will use another bank's solution to serve them ." And they say, "Okay, man, solve my problem, you understand?" Because if I do it internally, it takes me a year to have my solution implemented. So, the guy hands me all that stuff and that's what we do, right? So, the customers are liking it and that's how we see it evolving here, just like Fernando has done. And Fernando, what's he like? Taking advantage of the same line of question, right? When you started, did you always have this same proposal or did you understand along the way that there was a way to solve a problem that was sometimes even bigger than you imagined in the market? Dude, good question. We had an initial motivation which was to focus on vehicle recovery. Yes. But vehicle recovery in Brazil is quite complex, isn't it? Unlike the model that exists in the United States, where you put the car on the tow truck and take it away, right? There are a series of moves you need to orchestrate here, and it's not enough to just know that the vehicle is available in that location. You have several other steps: justice officer , search and arrest warrants, etc. There is a regulatory framework that is changing, but our reality still demands a lot of orchestration. And at first it was a point of frustration for us, saying " how difficult it is to gain traction in this," and at the beginning of the journey we ended up switching towards claims fraud prevention. So, it was an area where we moved out of the phase that we internally call "soft plaster," right? We shaped that data, that material we produced, fitting it into a value proposition, an offer, and a real market pain point. And that's how we focused on preventing claims fraud. So, understanding if that car had pre-existing damage, if there was a profile discrepancy; For example, if it was declared as private use but was used for commercial purposes. And when we bring up the subject matter of the image, we are talking about a real and visual fact of the world. You eliminate the declarative side of the information and the possibility of image manipulation, which is very fashionable today due to the capabilities of AI. So, for example, that inspection photo that the same person took, could it have been altered? Whether it's in a bank guarantee analysis process or with an insurance company. Could that image have been altered? Now, my images are captured in a completely passive way. The car is driving down the street and I'm recording that image. So there's no way that could have been altered during the process. So I know the actual condition of that property on that date, at that time, in that place. So this real-world evidence strongly contradicts that side of, for example, the postal code. It is declaratory. Do you actually live in that zip code, or are you committing subscription fraud by getting a better price on an insurance quote? Is the warranty being placed on the financing really in that condition ? Was that breakdown fixed with AI? So you have an alternative source of passively generated data, a 12-month history, which is where we achieved a very relevant success story and from there we started to move forward , looking for new sources of monetization for that data. And who do you primarily provide your solutions to? Today we have a strong focus on banks, insurance companies, and rental companies. We serve other markets, but these three are the main ones. In banks, for example, when formalizing financing or auto equity, you have a latitude and longitude of where it is being formalized, which today most operations are digital, you know where the store address is , but where is the asset on that date? I can mention this financing, for example, a real case we took, well, in Ribeirão Preto, a Fiat Mobi bought in Maranhão. It is very difficult to justify in terms of price the movement of a Fiat Mobi from Maranhão to Ribeirão Preto. So, a distance of more than 2000 km. That 's fraud. So there is an appearance of fraud in this case, it is a suspicion of fraud. And that's where we always recommend to the institution to carry out the verification. But it sounds very strange when you see that this triangulation is not harmonious, right? You have a guarantee that is very separate from the process that is being financed. So, that's one of the ways we help banks, in addition to the inspection part. So, for example, when you request an inspection, you create a lot of friction in the process. You depend on the person going to take the photo of the vehicle, going down, being in the parking lot, being underground, it being business hours, having time in their schedule. Then it creates quite a bit of friction along that route. Now, if I provide recent photos of the vehicle from all angles, the vehicle is in good condition and on the road, if you eliminate that need for the inspection link, then you have a frictionless digital journey for that process, you reduce journey abandonment, because when you need a user action in a digital journey, that's where the main gap in the journey lies. where there's room for something to go wrong, right? Yes, so that the person gives up, forgets, doesn't return, takes a long time to come back with the photos. Then you step out of that digital sphere and put it on hold. During that waiting period, there is a lot of abandonment of the journey, so when you eliminate that need, you create a fully automated flow. It's true. And, Lobo, taking advantage of the Carbid Data case discussed here , do you have any interesting cases you could bring up to help illustrate how you not only increased the collection effectiveness rate, but also any interesting details you think are worth sharing here for Falco? Ah yes , there are several, without a doubt. As I mentioned here, we have managed to make very significant reductions in terms of operations. So, for example, one of our clients, whom I even mentioned yesterday, had 120 people to handle collections with 2 million customers, and we implemented it and developed it with him, right? Today it has 10 million active cards. And it has six people to handle collections with an NPS of 73, right? So it's an evolution where, at the same point Fernando mentioned, you're observing the entire journey, right? This is done directly with AI , right? You're observing the entire journey called the performance table, which is digital with human input, understanding where you have the most friction and how to clean up those frictions. And sometimes small nuances that seem to have no impact , have a big impact. For example, this same client said: "Look, it's good, if the client takes 10 minutes, close the session and reopen it, for example, for payment." "For that reason alone, the ability to collect payments fell by half ." Nobody knew what that "log out" point was , right? So, you can understand what the critical points are through a dynamic assessment, right? The moment you go from 10 minutes to 1 hour keeping the client connected, you double the recovery volume, right? The same applies to the CPF. The customer is very afraid to enter their CPF number. If you have to enter your CPF number, you're in for a huge loss, almost half, because you're afraid it's phishing, some kind of fraud, you know? So, alternatives of this nature, like the one we made, can duplicate each point of that funnel and thus greatly increase effectiveness. So, this analysis is done online, and we manage to improve the company's efficiency daily at an absurdly low cost. Incredible. And even Bruno, if I may add another comment, I think we talked about friction in the journey, which is always a big issue, because you're eliminating flow opportunities in that journey that we invest so much in to keep it in . And I think another angle that we're also pushing a lot at Carat is about hyper-personalization, right? Moving away from this average risk, average price view, right? Stop looking at the average of a group or a segmentation and hyper-personalize, for example, in our case, by enrollment. In their case, it's probably the person, the CPF, the individual, but in our case it's the registration number. So, for example, it's almost the same dynamic that existed in the financial market with the change from a negative credit bureau to a positive one. Previously, we set prices by looking at negatives and created an average risk that the score read, but with the positive bureau, you understand the payment behavior of each CPF individually and personally. It's similar with the license plate. For example, what constitutes a personal use? What is considered commercial use today within the context of Uber? For example, there is a collaborator of ours who works, would declare private use calmly, employee, working in the office, but on returning home decides to do Uber, decides to make some logistics deliveries. Then you ask yourself: "Wow, that role is getting confused, is it worth setting the price for what is a particular use?" What is commercial use? What is a car in this category? What is a region in this category? I need to look at that license plate and understand its status, where it's being driven. So that hyper-personalized look, I want to look at the license plate and give a price for that license plate and no longer say that the average profile of this region and of this car is X in terms of risk or pricing. Things are no longer so binary, and we have the tools to capture those nuances we see in different markets, right? So that's great, no doubt about it. And I think Fernando's point here is very good, don't you? Increasingly, companies that manage to achieve that individualization will bring back to the table the lost money from that average, right?, in which the vast majority still operates. The moment you manage to individualize, and that's the path of those using cutting-edge technology , you manage to bring margins back to the table that were being lost due to inefficiency, and you give that back to our society, right? Because it costs less, doesn't it? More optimized, cheaper credit, cheaper insurance, more people in the credit market, more cars covered with insurance. So it's that mechanism we need to make work better. And I believe that both Fintx and Car Big Data have been bringing these opportunities and tools to make possible the reduction of friction, the reduction of loss , optimization, automation and this hyper-personalization. For me, these are the key elements to make this machine work better and reap those benefits even as a society, right? not only as a company. And another thing you have in common, obviously we have what becomes normal and widespread with the use of artificial intelligence and that refinement of all this, but first of all you are companies that are very focused on data also within your specific markets, right ? So, on one hand, we have here the part of conversation, voice, context, intention. And in your case here, data, circulation, location, behavior of assets, right?, in the physical world. And so, at the same time as we see this data influencing decisions, it also feeds into the very AI models we've been talking about, right? And we are seeing an increasing concern about the quality, privacy, security, and explainability of the models of what is being done. How do you think it's possible to balance the potential to create proprietary intelligence from all this—which is what you do with that data—with the need to also build trust and governance around that data, within the framework of everything that is advocated in terms of good practices and also the law, right? I think it could be. Yes, that's quite interesting. When we were born, I came from Itaú, and the design was as follows: we have to have Itaú and Bradesco. If we manage to get past the Itaú and Bradesco filter, the rest will fall into place , right? We have both of them here in our portfolio. Anyway, we already work with them and all the design we did was for the banking segment, which has, well, quite rigid legislation, right? Fairly strict requirements with penetration testing all the time, as well as encryption precepts at every level. But there's a point here that I think is important, which is the point that Fernando mentioned, which is: you want to keep that individualization, you want to use that behavior, but you can't share it with everyone. Exact. But you have a behavior , and by knowing that behavior—and obviously you're always there, there's no way nowadays not to be aligned with data protection laws and all legislation—you manage to help the customer have better conditions, whether it's... Whether it's the customer who has a car, therefore, they can have cheaper financing , cheaper insurance, really , which they couldn't get because they would be falling into an average, right? as well as helping the company to have a more efficient model. The point you made is quite interesting, and so the good thing about AI is that it allows many companies to come in and do great things. The downside of AI is that it allows many companies to share information they shouldn't be sharing, right? Partly because there are many people who are not prepared, right? I think the quality of AI is the quality of your ability to discern how you use it, right? Unfortunately, not everyone has the experience and discernment to use this in the best possible way, mainly due to concerns about security, right? So I still see a lot of problems ahead, a lot of problems happening with, well , businesses that aren't ready to use this. For example, everything we use related to anything AI-related, first of all, are closed models, right?, guaranteed by the providers, but in addition to that, everything is encrypted when it comes to sensitive data. So, no sensitive data is passed on, it's masked, so the AI ​​never has access to any data, right? But not all companies have that care and concern. And how have you dealt with this? Yes, nowadays, when we talk about images in public environments, we take the LGPD very seriously. So, this stems from the origin of the company and the discussion. So from the beginning we were focused on that. Therefore, the regulatory vision was there from the company's inception ; It was not an adaptation at an opportune moment. And today, for example, we are forced to perform facial recognition on images to erase people's faces on the street. Ah, so at my base, for example, I don't have any faces, I only have vehicles, right? And that is an obligation we have due to a rule of the LGPD. And we also have very strong governance of the use case, right?, regarding how that data will be used, what the purpose of using this information is, right? So, for example, in the relationship with the bank it is the bank's guarantee, in the relationship with the insurer it is the insured vehicle, there is a contract execution between the parties. So, there are limits within which we can act, and we greatly respect those limits and the responsibilities we assume. And I think, as Lobô was very right in his comment, every startup jokes, right? If I was approved by these people, then we are on the right path to governance. So, wow, today we're working with the biggest insurance companies in the country, the biggest banks in the country. They are rigid processes that force us to always improve. But I think one point where we are much more mature in the market, which I observe, is that these conversations began at the origin of development and not at a later moment of adaptation. So, I think Brazil has already matured a lot in terms of privacy, bringing this conversation into the present and being born within regulation, right? A point was mentioned here that I want to address to both of you, and that is that you have managed to act and work today with very large clients, with very high standards, right? What has this journey been like? To what do you attribute this success, this approach built from the beginning? It's an adaptation that you develop as you interact with them, because the solution is very interesting for this audience, but obviously it has an incredible aspect. I even wanted to share this as a learning experience for whoever is listening here, who is also on this journey of trying to sell to a tough corporation, right? The question is for both of you, we'll take turns. Lou, go ahead. Thank you. Well, let me tell you, selling to a big company is difficult . I think that's very difficult, isn't it? Uh, they're long cycles, aren't they? We're talking here, I don't know, an average of 9 months, right? We're going from six to one year, right? the cycles. Six is ​​a case , isn't it? Because you have to go through 500 filters, 500 is really complex, is n't it? So, you need a lot of patience and emotional resilience, because the process has its ups and downs, right? They take vacations, then the other area takes them, it's a long process, isn't it? And but it's interesting because, for example, for us, we decided that our marketing was about having strong logos, because, right?, if you're in a big company A, B, C, that already endorses you, right? It already vouches for you. But ? So, our original design was, man, we're going for the big logos, right? It will be difficult, but once we have them, then that was it, that was strategically planned. Ah, yes. Design, it was by design, wasn't it? Obviously you need to have breath because until then there is no income, the whole thing is complicated, is n't it? We feel very tempted to share this with those who are just starting out . We were very tempted to acquire smaller companies, and it's interesting; we had several large companies that would have provided good revenue, but we did n't acquire them because, as with any startup, we have limited capacity and have to make choices. Yes. And we prefer to choose large, well-known companies , even if they pay less. Ah, to get the seal, because they leveraged it. So, it's much easier to sell today than it was in the past, right? Where are you located? I 'm nowhere, wow. Selling is crazy, isn't it? So where are you? I'm in those companies, man. Excellent. You're doing well, you're well positioned. Halfway there. Halfway there. Now it's a difficult choice. It's a difficult choice, isn't it? I'm sure life wasn't so easy for Fernando here either. It 's never easy. And that's how it is, there's a lot of bitterness because there are back and forths, planned times that are not met. So the entrepreneur suffers a lot on this journey, because we are anxious to see the result happen . No, and Brazil is peaceful, right? There are no surprises in this country, regulatory, there's nothing, right? Business atmosphere and all , right, buddy? But, but I think that, an important fact, which I think is this choice of strategy of big logos, right?, blue chips there that are going to help us create more traction in the market, I think it's a very relevant strategy and it's a choice that we have to make, and I think Lobo was very right. Dude, it's an election , at the end of the day it's an election. We want to serve everyone, but we have to choose because capacities are limited and we have execution commitments that we have to deliver, right? And I think the balance between the innovation agenda and the results agenda speaks volumes about this, because many times, wow, there are a lot of interesting talks, a lot of great ideas, a lot of things that, wow, it's nice to talk about, create, innovate, think outside the box, but we also need to invest, because it's a business that needs to grow. Our growth rate needs to be differentiated. So I need a results-oriented agenda, but my characteristic is an innovation-oriented agenda. So I believe that maintaining a balance between innovation agendas and results agendas are the choices that allow us to have a good trajectory. And this positive balance usually comes from large accounts, because even some time ago, large companies were also somewhat averse to giving opportunities to very small companies. That changed the mindset a lot. That changed that mindset a lot. So you see, right?, various parts of companies, structures dedicated to innovation within the organization itself. Therefore, even reducing internal bureaucracy to achieve faster framing . In the past, people asked for too many things that we were still asking for, but so it improved, it improved, I think in view of what already existed. It's because the day of the person who's there in those shoes, I know , is more complicated. Yes, the day is tough, but I think the receptiveness improved, the start of the conversation improved, and then you have a long day until you manage to materialize that. In the past, I think, depending on the level you were dealing with, at a strategic level, they always wanted innovation, but depending on the level of the company, you would encounter resistance due to fear. Oh, yes, right? Now I think there isn't one; people are realizing there's no option. Adopt or die. And there are innovations as specific as what you do, which also weren't going to come from within the institution, right? You know what? There's no way a large company can innovate at the level we can, because how can you use all the models and test them in a large company if the governance doesn't allow it? Right? I had a huge consulting firm I was talking to and they said, "No, Lobo, you know that if I want to , I can make a platform like yours, right?" I told him, "Dude, you know you'll never be able to do it because you don't have the governance that allows you to do it." So, it's not that they don't want to or that they don't have good people, but the structure prevents innovation, prevents them from having the freedom to do a number of things that we do have. So we managed to bring in, we managed to be an innovation arm that the bank will never be able to reach at the level of sophistication that Fernando and his team have. Never. There's no way. Besides some defensive moats they have here, like the camera issue and the association, where you're going beyond what we talked about today ; We build faster with AI, for example, but there are externalities here that are not at its core and it will be very difficult to make them interested. And a very fast-paced approach, because I have a focus, I have a hyper-focus on what I do, I'm going to be faster and more efficient in what I do. Capacity to do, well, any bank has the capacity to do whatever it wants, right?, in terms of resources. But the challenge lies in agendas, focus, the core business, and that slows things down, and we, as startups, tend to narrow our focus, right? So that approach gives us room to have that agility even with a lot of governance too, because we are agile, but we also self-govern, right? We also created that, otherwise, brilliant agility is useless if it comes time to hire. No, no, I'm not certified. So that balance also comes a lot from the approach. We are focused on what we do. Alright, everyone. We're reaching the end here. I wanted to hear from you all, to wrap things up, right? After winning a competition like this, what is the next challenge you would like to have overcome when we talk again in a year, for example? And also, what is your assessment of these three days here in the lounge? Uh, this is the first time for both of you, right? If they could do that closure here too, I think that would be great. Well, that's the challenge , I wanted to be present in every place where there's a plaque , right? That's the big dream, that's the dream. But jokes aside, it's really about understanding that the market won with this hyper-personalization, this individualization, that we managed to transform that into results, that we're coming on this journey, and I think it depends a lot on market acceptance for this to happen faster and for us to bring it to present value. In terms of innovation, I think that we, as a company, have a strong focus on image, and it's natural that we start looking at video. So that's also a boundary we hope to break very soon, of saying, why ask for an investigation if I can have the video of that accident happening, press play and watch it, for example? So, there's an innovation frontier we're also looking for, which is to continue along this multimodal path and always learn more about the car, right? Our appetite for vehicle data isn't going to decrease, so we want to know everything about a car. We are on this trajectory in terms of space here. I think it's really cool. I think the dynamic is incredible, getting to know all the business partners that are being formed and this exchange is very enriching. So I think this space for debate at the biggest financial market event in Latin America, if I'm not mistaken or if it's not bigger than that, there's no doubt. Yes, it's too good. So, being here is an honor, a pleasure, and the challenge is to be here more often, right? Whether it's with these businesses or with new ideas that we're going to create. Let's go. This is just the beginning of this journey, my dear Wolf. Tell me. Look, our dream is to improve life . of people. We are focused on collections, but our focus is very much on the person. We're in a country with an insane level of debt, right? Uh, I think more than half the country has some outstanding debt. And our vision and mission is how to help these people reintegrate into the financial world, to get through this delicate moment in their lives as smoothly as possible, right? Right? And what we are doing, and what we believe in, and the path we are taking is the integration of countless types of payment methods and conditions to change people's debt profiles, sometimes, usually , moving from unsecured loans to secured loans and different solutions, so that these people can pay smaller amounts, under conditions that suit their budget, so that they can get their lives back on track. So our mindset is to help Brazilians regain a positive position in the financial world, right? And that's very important because the moment you have financial independence, you can dream again. Dreaming of paying for your child's education, dreaming of buying a car, dreaming of owning a house, but you free yourself, you need that. And at the same time, by doing this, we solve the problems of businesses. It's true. So, the design we have and the mindset we would love to apply is to help Brazil solve this debt problem in the easiest way possible and with the least pain and suffering . And that is our mission. We're loving being here. This exchange is wonderful. I think this event is amazing. In fact, I have some here, just to share with the team, right? I think all of us here go to fairs here and there, etc. I was commenting here, I think this is the most organized fair I 've attended lately. I was at several trade shows in the United States last year and such. Wow, this is amazing, isn't it? The entry is painless. You enter here, there's no problem logging in with facial recognition. The fair is working incredibly well. I'm very impressed, aren't I? I think this is perhaps the most amazing fair, right? Besides being number one in Latin America for financial services, you'll find everyone you've ever met here. It's beautiful, I'm loving it. Yes, this is a very good meeting point . And friends, I wanted to thank you enormously for your participation here and also congratulate you on the results. I hope they did good business and that they'll be back next time . A round of applause here for our finalists and also for the panelists. And that's all, folks. See you next year, then. It's excellent. Thank you so much. A big hug.