Building Trust in Fintech | "What The Fraud?" Podcast—Money 20/20 Special Edition, Part 2

For part two of this special episode of What The Fraud?, Sumsub’s Sales Director for the Americas, Anastasia Shvechkova, sits down with five industry leaders who are scaling trust without slowing growth. Fraud is getting faster—so strategies must get smarter. What are the leaders who rethink risk doing to stay ahead?

Recorded live at Money20/20 USA, this episode features insights from experts at Standard Chartered, EBANX, ACAMS, Adyen, and Nubank. Together, we unpack the modernization of dollar infrastructure, approaches to combating fraud in emerging markets, the evolution of global AML standards, the challenge of breaking down data silos, the realities of Web3 compliance, the management of BNPL risk, and the growing influence of AI in global payments.

Jennifer Lassiter, Standard Chartered

ANASTASIA SHVECHKOVA: I'm joined here by Jennifer Lassiter, Head of Digital Assets for Europe and Americas at Standard Chartered. Hi, Jennifer. 

JENNIFER LASSITER: Hello.

ANASTASIA SHVECHKOVA: How has Money20/20 been for you so far?

JENNIFER LASSITER: Fun. Lots of energy. Great momentum running from place to place, but really valuable and to the point conversations.

ANASTASIA SHVECHKOVA: That makes a lot of sense. I know that you had a panel, so your panel at Money20/20 is titled ‘Defending the Dollar’, where you discuss how the US dollar, while still the backbone of global finance, is being held back by outdated infrastructure. So, I wanted to kind of ask about that. Why do you think the dollar payment rails feel outdated today? And what would be a better version actually look like?

JENNIFER LASSITER: This is a really important question. A lot of people ask it because the user experience in the US, in particular, is pretty instant and social when it comes to payments. However, what's happening on the back end has not been modernized, and it is still reliant on existing banking systems to support clearing and settlement, as well as the movement of money. So, when we're talking about cryptocurrency, blockchain systems, etc., what we're talking about is changing that back end and creating an instant social and seamless experience in the movement of that money to match the front-end experience that people have today. What is very interesting from a Standard Chartered perspective is that we've been able to do that in specific regions to date. And what we're looking for, as we see the momentum emerging from the US, is connecting those cross-border pieces so that the technology on the back end enables the transfer of value from region to region.

ANASTASIA SHVECHKOVA: Yeah, I think with all the stablecoin movement right now, it's like the best time to help modernize in general. Right?

JENNIFER LASSITER: That's right.

ANASTASIA SHVECHKOVA: How do you balance a bank's need for stability and compliance with the agility that comes from things like tokenized or programmable deposits?

JENNIFER LASSITER: I like this question, and I hesitate only because I don't think it's an 'either/or'. And that's a really important distinction. And for Standard Chartered, the way that we've been able to grow our global network and our digital asset capabilities is by being very intentional and learning through different vehicles, how we create an institutional-grade risk framework and meet the compliance requirements in the respective regions we're operating in.

What I mean by that is, like many other banks, we have a ventures arm, our ventures arm, we really look towards being able to say, 'Okay, we've got a situation in the market that we know needs to be addressed.' New product or client demand. Are the solutions that are provided commercially viable? If they're commercially viable solutions, are we able to put a risk wrapper around it and then be able to bring that learned experience into the core bank offerings? Sometimes we do that and sometimes we don't, which is a learning lesson for us. As it stands, the way that we address digital assets across our global network is actually a business-as-usual activity. And what that means is that we don't have crypto experts who are necessarily sitting in an innovation division. We have crypto experts who are helping to inform global strategy and make sure that our regions have the tools they need to be able to execute. However, the actual risk lies with the product manager for the digital asset capability within the business, and that means we have had to implement risk and compliance frameworks and governance across all of those business units. So, back to your original question. I think there's healthy friction, and we have to pay attention to the risks associated with it, but it has not varied for us in terms of how we address risk and compliance across the bank historically.

ANASTASIA SHVECHKOVA: Got it. And with this versatility, let's say in different ways you do that and modernize, which challenges do you see when it comes to compliance and fraud? And in general, I think, this kind of speed of innovation.

JENNIFER LASSITER: I think it's important, and a conversation around fraud, to call out that the technology itself is not inherently ripe for fraud.

Where you have money, you will have fraud.

And it is a very important and critical thing for Standard Chartered to pay attention to and to understand how to address and where there is uniqueness in blockchain as a technology.

You mentioned speed as one factor that helps us understand how to address it specifically. Additionally, in that question, I would bifurcate between what are traditional banking activities that we're doing to support a new client segment. And that can be banks that touch crypto or stablecoin issuers or operators—but it's still traditional banking services that we're providing them. And so, the traditional risk and compliance frameworks would apply.

Then there's a second category where we might actually have Web3 native products or be working directly with a client that has those products, and there's a blockchain nexus. We also need to understand the fraud and compliance requirements associated with that particular banking activity. And when that happens, we've had great success in finding partnerships with global compliance institutions who do this day in, day out, and are helping us monitor that activity on-chain and off-chain.

And then finally, I would say we've been very successful also in the regions where we have legally permissible activity, working with the regulators to help provide thought leadership around how compliance and risk management can unfold with these digital asset capabilities. I was really excited when the New York DFS released its blockchain industry guidance last month. I guess September, maybe that was two months ago now to be able to point to that and say as a bank that we are confident that we're meeting the blockchain guidance that's already been put forward and being able then to take our global experience and bring that into the US to continue to inform and provide thought leadership on how that can be done.

ANASTASIA SHVECHKOVA: I wanted to piggyback on what you said about Web3 because we haven't talked about this yet. What is the biggest difference between the infrastructure of compliance and making sure that you protect the customer and protect the business in Web3 versus Web2 or traditional finance?

JENNIFER LASSITER: I think it depends on the activity that you're supporting, and there are nuances to whether or not we are doing something, such as providing a settlement function that includes a minting and burning of a stablecoin, for example. That's going to have a nexus with blockchain that just requires additional transparency and auditability in the reporting of that activity. For the most part, the activities that the bank is participating in today, it's the traditional banking components for our crypto clients. And it's providing access into the markets, whether it's crypto as an investment or using stablecoin for a corporate treasury purpose. And those are traditional activities that we look at from a risk and governance perspective, that's on par historically with our banking activities.

ANASTASIA SHVECHKOVA: Perfect. I have the last one for you, coming back to the dollar infrastructure. So, looking ahead, if we want this next generation of dollar infrastructure to actually work, and not just technologically but safely, what do you think banks, fintechs, and regulators need to get right now?

JENNIFER LASSITER: Okay, I'm really excited about this question. And the reason I'm excited is because in your question is also the answer. And I think it's a bit unique. Right now, the conversation folks are coming to, 'This is my swim lane. This is what I need to have a bank be successful.'

I actually think for the ecosystem to be successful and to see long-term and sustainable change in a safe and sound way, you need all of those stakeholders at the table. You need fintechs, you need banks, you need native crypto operators. You need governments.

We all need to be working in a very orchestrated way together to determine what the outcome is that we're trying to achieve. The reason that's so critical is that the use cases we see today, in my opinion, are not a limited set of use cases.

As technology scales, we will see new use cases emerge, and we need to have regulations, legislation, and industry-set standards—all of which allow for flexibility to accommodate that scaling and these new use cases. That doesn't happen if there's only one stakeholder participating in the conversation.

So, I think working through trade organizations, making sure that you're providing thought leadership back to governments and participating in public-private opportunities, is really critical to getting the guardrails right for a safe and sound ecosystem in the future.

ANASTASIA SHVECHKOVA: That's a great point. We talked here about the importance of collaboration within the company, but now you're talking also about collaboration outside of the company and talking to governments. So, this is brilliant. Thank you very much. Very nice having you here. Enjoy Money20/20

JENNIFER LASSITER: I will, you too.

ANASTASIA SHVECHKOVA: Thank you. 

André Peixoto, EBANX

ANASTASIA SHVECHKOVA: I'm joined here by André Peixoto, Director of Operations at EBANX. André, welcome.

ANDRÉ PEIXOTO: Thank you. Thank you for having me, Anastasia.

ANASTASIA SHVECHKOVA: Yeah. I'm going to have some questions for you around your models and fraud prevention. But let's start with talking about scaling a payment platform across so many countries and what it requires for merchants and consumers and how much trust you have to build. What internal systems or operational principles does EBANX rely on to maintain performance and reliability at that scale?

ANDRÉ PEIXOTO: Great question. EBANX today processes over 3 million transactions every day, and we rely on two main pillars to guarantee scalability and performance. The first one is technology. So, we have a unified API, a layered approach to fraud prevention, and everything is cloud-based. And we rely also on culture being performance-obsessed. So, we're making sure that all teams inside EBANX are really accountable for performance, from engineering to product to operations, which is my area. Customer success. Everyone is extremely focused on delivering the best performance. So, we hang on to processes and scale our infrastructure to ensure that we deliver the best every day.

ANASTASIA SHVECHKOVA: We talked a lot about the connection between teams today, actually, so that really speaks to what other people experience and what they vouch for. But from your experience specifically in South America, let's say, how do issuing banks approach risk, and how did you have to adjust your strategy accordingly?

ANDRÉ PEIXOTO: In South America most issuers tend to be very-very risk conservative. So, they employ different measures to control risk like dynamic CVVs, which is like refreshing CVVs inside the apps or what they call the botónera, which is enabling or disabling the card for e-commerce transactions. And this, of course, takes a heavy hit on the authorization performance. They manage to control risk, they manage to safeguard the ecosystem. But at what cost? The merchant and the customer in the end suffer from degradation, in authorization rates. What we do is have a very, very close relationship with those issuers. We're talking to them, like sometimes on a biweekly monthly basis to understand what's going on, what are the fraud trends, how EBANX and them and the merchants can work together to guarantee we're minimizing risk while also maintaining a healthy approval rate.

ANASTASIA SHVECHKOVA: And what are the challenges? So, you mentioned the balance between the conversion rate, of course, and the fraud prevention. But what do you think right now keeps you up at night, let's say when it comes to fraud. I know that in Latin America, fraud is very innovative, let's say. So, what are the fraud patterns? Let's say that you see.

ANDRÉ PEIXOTO: The buzzword 'AI', right?

ANASTASIA SHVECHKOVA: Yeah.

ANDRÉ PEIXOTO: That's what fraud has been very sophisticated in using AI, using fake storefronts, using synthetic identities, triangulation fraud, doing inventory hoarding. A lot of these mechanisms that fraudsters are making very good use of generative AI to employ, and that we need to also be very well-prepared with AI and different strategies, different tools to fight fire with fire, let’s say.

ANASTASIA SHVECHKOVA: And which kind of models maybe you build and maintain internally, and how much AI do you use, and how do you approach this responsibly?

ANDRÉ PEIXOTO: We have tools, internal AI models, that are using like 1000 rules to score the transactions, understand really the behavior, the patterns, to guarantee that we see the transaction score by what it is: the predictability on whether it is a fraud or not, and taking decisions on top of that. Besides our own tools, we also count on global providers like names and big names in the market like Riskfied and Signifyd. And we unite those strategies, our own internal tools, our own internal AI models with external providers as well, to maximize the efficiency of the fraud engine.

ANASTASIA SHVECHKOVA: Great. I think it's important to have strong partners in that fraud prevention. And the last one. But looking ahead, where do you see the biggest shifts coming in payments across emerging markets in 2026, whether it's local methods, infrastructure or technology, and what can redefine access and inclusion?

ANDRÉ PEIXOTO: I think the first one is APMs becoming instant, becoming real-time, becoming frictionless, and becoming even recurring. Start with UPI, which set a benchmark for the whole world. Pix in Brazil followed that up very closely. Now we have Bre-B in Colombia. All of these are peer-to-peer instant APMs that also offer recurring capabilities, allowing subscription merchants to take advantage of this feature and provide access beyond card users.

And we also have digital wallets offering the same: recurring capabilities, less risk. So, I think the big trend, number one, is these APMs becoming more instant, becoming recurring, and giving more access to customers.

Second one, I would say, stablecoins. So stablecoins have evolved from speculative assets to really a payment instrument, allowing for real-time settlements, allowing for instant payments, and really enabling even more access to customers and facilities to merchants. This one has been growing. EBANX is already digging into that as well, offering stablecoins as part of the payment flow, a part of the settlement flow.

And lastly, AI—also the buzzword. EBANX has also been employing this approach with its anti-fraud model. But also, we have our performance engine based on AI models as well, which consumes the score and understands what's the most likely path for that transaction to be approved. We work with various MIDs on acquisitions and route them accordingly to understand the risk and the vertical. Based on this understanding, we achieve up to 14% higher approval rates using our AI-based routeing model and the fraud prevention model for the Xshield.

ANASTASIA SHVECHKOVA: Thank you very much. I think that, like a lot of people here at Money20/20 would agree with those three, actually, because those are buzzwords, but they are important because that's what's happening. That's the world around us. Thank you very much for joining us today and for sharing your thoughts.

ANDRÉ PEIXOTO: Thank you. Thank you for having me. It's a pleasure.

Craig Timm, ACAMS

ANASTASIA SHVECHKOVA: I'm joined here by Craig Timm, Senior Director of AML at ACAMS, one of the world's leading organizations for financial crime prevention. Craig, welcome. It's great to have you on.

CRAIG TIMM: Thanks so much for having me.

ANASTASIA SHVECHKOVA: Well, let's start with maybe my first question about AML teams. They have gone through a huge transformation in the past few years. How have you seen that role evolve across financial institutions?

CRAIG TIMM: Yeah, I think we've seen what seems like a never ending, increasing set of expectations that comes in a few different ways. So, one of just the portfolio of work. So, if you think about how fraud has evolved, it is more focused on fraud against customers or sanctions evasion. The traditional tools we have to fight fraud or sanction screening don't necessarily work as well there.

What does work well are AML tools and AML investigative skill sets.

So, a lot of those activities are now falling on AML teams. Then, if you look at things like data and technology and all the expectations around that that are falling on AML teams, the skill sets that you have to have the people, the volume of work has just really increased.

ANASTASIA SHVECHKOVA: I think a lot of our listeners, of course, are fighting fraud every day. And this increasing responsibility that they have is crazy. But fraud and AML teams often operate in parallel. From your experience, what's the most effective way to get fraud and AML teams working together in actual practice?

CRAIG TIMM: Yeah, it's one of the biggest challenges to detecting this activity earlier to being more effective. I'd say it's something that everybody's looking at. I don't think there's a one-size-fits-all, but I think there are some common principles that people are looking at, and it all sort of centers around looking at your people, your processes, and your technology. So, people, it sort of runs the spectrum from some are just bringing their orgs together. They're saying that's how we're going to do it.

Short of that, others are co-locating teams or creating new committees or just any way that we can bring these groups and the people, cross-training closer together. Around processes, it's looking at your policies and procedures. Are they consistent where they should be? Do they clearly lay out roles and responsibilities, handoffs between the teams—all those sorts of things. And then technology, just like the teams are siloed, the tech is often siloed. And it's like, hey, could we share some of this data? Should we use common tools? Should we be on the same platform, all in an effort to bring better context to help detect the activity and ultimately report better suspicious activity to the government?

ANASTASIA SHVECHKOVA: Exactly. And when you say effective fin-crime prevention, what does it really mean in practice in 2025? Because it's a lot of different things, as you've mentioned, in AML and fraud prevention and so on and so forth. What does it mean exactly?

CRAIG TIMM: Well, I think we're at a really exciting time, particularly in the US, because I think it's changing. Like, if you think about anti-financial crime, the whole goal is to make a difference in fighting crime. Now, that's not been how these programs have been examined for a long time, but that's changing. So, the Undersecretary of the Treasury recently gave a speech where he talked about a transformation of the AML regime, on whether we are actually producing effective information for law enforcement. And he's talking about reorienting the regime around priorities, changing the examination process. So, I think it's a really powerful moment in time where we could get to a point where we define effectiveness as whether we are doing a good job. And then that's how we evaluate financial institutions and others. It just could have a transformative effect and make our communities much safer, frankly.

ANASTASIA SHVECHKOVA: Going back to that AML role question at ACAMS, you train compliance professionals all over the world. Which skills, mindsets or cross-disciplinary knowledge do you think risk teams need most right now to stay actually effective?

CRAIG TIMM: Yeah. So, I mean, I think the biggest thing I hear from everyone all over the world is just the scale and pace of change is overwhelming. And so, you need the skill sets to be able to handle that. So, I think adaptability is one. I think you got to really like learning. When you look at all the regulatory change and you look at the way criminals evolve, if you don't wake up every morning excited to learn about the latest techniques, you're just going to fall behind.

And then the last thing is tech. AI will transform what we do in fighting financial crime. And it's something that everybody's not going to have to be a data scientist. So, we're hiring a lot more of those. But everyone's going to have to understand the tech. They're going to have to know how to use it, and how to deploy it effectively. So, I think that's another thing that maybe five years ago, somebody that's a lawyer like me might not needed to have those skill sets. But now pretty much everybody does.

ANASTASIA SHVECHKOVA: Definitely. We had Sumsub also say that it's an arms race out there, right, the fraudsters evolve, we evolve, and the tech evolves. It's a never-ending cycle.

CRAIG TIMM: And the criminals are winning right now. And we have to fight back. If you look at the fraud numbers, they're winning. We have to fight back.

ANASTASIA SHVECHKOVA: That's true. I have the last one for you. Looking ahead, what trends or innovations do you think will most reshape AML and financial crime prevention in 2026?

CRAIG TIMM: It's two things. It's the focus on effectiveness. And can we actually change our programs to focus on results?

If we can get resources away from process and checking boxes to actually things that matter, that could have a transformative effect. And then two is AI. We're here at Money20/20, I think one of the biggest buzzwords has been agentic AI. I heard the CEO from BNY talk about digital employees. And I heard somebody else say, by 2030, there's going to be 6 billion AI agents. This is going to transform every aspect of our lives, including financial crime. And I think it could have a really positive effect. So, I think it's pretty exciting.

ANASTASIA SHVECHKOVA: I agree. Thank you very much, Craig, for your time, and hope you enjoy Money20/20.

CRAIG TIMM: Great. Thanks so much for having me.

ANASTASIA SHVECHKOVA: Thank you.

Brigette Korney, Adyen 

ANASTASIA SHVECHKOVA: I'm here with Brigette Korney, Global Head of Risk at Adyen. Brigette, welcome. Thank you very much for joining us today.

BRIGETTE KORNEY: Thank you for having me. Nice to meet you.

ANASTASIA SHVECHKOVA: How has Money20/20 been for you?

BRIGETTE KORNEY: Crazy per usual, but exciting. I think there's been so much energy throughout the whole entire show so far, and even just walking through, you can really feel how happy people are to be here.

ANASTASIA SHVECHKOVA: Yeah. It's happy. It's great to connect with everybody. We're also happy to be here. I'm going to start with just questions around the nature of risk leadership and how it has changed in that environment where fraud tactics and regulations evolve at lightning speed.

BRIGETTE KORNEY: Fraud is a game of whack-a-mole. I think that is the common term that people use. Risk leadership and fraud in that space has evolved just as fraud techniques have evolved. We've had to really stay on our toes. And I think as a fraud prevention expert, typically you'd think that we are risk-averse and that we're really there to stop fraud versus being maybe a little bit more out of the box thinking. And so, I think one of the best parts about the evolution of risk leadership so far is that we've been able to be really creative in how we solve and stop fraud across our customers. And my background is in customer service. So, in that space, we were really focused on making sure that the experience of the customer is positive. And we're not just stopping transactions and blocking, but we're really being smart about how we look at allowing positive transactions through.

ANASTASIA SHVECHKOVA: And what are the core principles that help you to actually balance this kind of innovation and growth with this risk management, as you said, like thinking outside of the box?

BRIGETTE KORNEY: I think that just, kind of like what I was talking about, is looking holistically at the entire transaction flow, really caring about what the friction is that the customer is experiencing on the front end when they go to press buy. What is that experience like for them? Authentication? Is there visible tokenization to them? Do they need to be involved in any of that process through either authenticating their bank account if you're using 3D secure and then all the way through any sort of fraud prevention that you're doing, making sure that that is balanced with the entire experience of the customer and so that the whole transaction is really seamless and that you are you're really balancing both sides of fraud prevention, but also increasing and improving performance as much as you can.

ANASTASIA SHVECHKOVA: Great. And I think one of the biggest themes here at the conference is breaking down data silos. How does Adyen actually leverage its unified platform and data visibility to detect fraud signals earlier, faster, to protect the whole life cycle of the customer? And what do you think is still missing in terms of cross-industry collaboration?

BRIGETTE KORNEY: We talk so much at Adyen about data. It is a constant topic because one of our core principles is that your fraud prevention techniques are only as strong as the amount of data that you're using to stop them. And so, the Adyen platform has over a billion unique identities. That amount of data alone helps us guide our customers to be able to stop as much fraud as possible.

ANASTASIA SHVECHKOVA: Brilliant. And I wanted to also ask about AI because this is like the hype word, right? We're talking to everybody about this. And, as fraudsters increasingly harness AI in more sophisticated ways, the fight against financial crime is becoming pretty hard, right?

BRIGETTE KORNEY: Yeah, absolutely.

ANASTASIA SHVECHKOVA: How can you ensure that AI remains a force for good against this kind of backdrop?

BRIGETTE KORNEY: I think we've talked a lot about balance so far today, but I would go back to balancing using AI techniques with the element of a human still, and not just completely moving away and moving towards a machine. If we can keep an element of how a person thinks, I think that will help us make sure that the AI techniques that we're using are actually effective and at the peak that they can be. Fraudsters are also using AI. It goes on both sides. And so, making sure that as we create more and more fraud prevention techniques that leverage AI, we're just not completely eliminating humans from the conversation, so that we can keep smartly thinking about it.

ANASTASIA SHVECHKOVA: And what would you say is the biggest challenge right now that you face in terms of fraud or risk? What has changed in the last few years, or what has been new apart from AI? Because that's given. But like anything else that you're seeing that is kind of threatening.

BRIGETTE KORNEY: Being in this space, I feel like a lot of things keep me up at night. Number one though other than Agentic AI or how we evolved there, I think working together is the thing that is really like making sure that we are focused on creating a great community for fraud prevention. I guess you could call us fraud fighters, making sure that we're all leveraging the same intelligence across each other and making sure that we're not letting the fraudsters win by. They have their own communities as well. And so, how do we make sure that we are not losing sight of what the endgame is? That's creating the best customer journey that we possibly can. That's improving authorization to the highest amounts. But also making sure that we're doing that as frictionless as possible and in safe ways. And so, how do we do that? As always, it's going to keep me up at night, has forever. But I'll stick with that one.

ANASTASIA SHVECHKOVA: It's like to fight the fraud rings. You got to create your own ring of fraud firefighters. Right?

BRIGETTE KORNEY: Exactly. We all have to be in it together.

ANASTASIA SHVECHKOVA: Yeah. And the last one, I guess. Looking to the future, how does the way we approach fraud and risk need to change from your perspective?

BRIGETTE KORNEY: I think it goes back to what I was just saying. We need to ensure that we are all working towards figuring out what's best for the customer. That is, the customer has to come first in that conversation. And how we do that is, is through creating a community.

ANASTASIA SHVECHKOVA: That's great. Brigette, thank you very much for being here and enjoy. Money20/20.

BRIGETTE KORNEY: It was so lovely to chat with you. Thank you so much.

Nathan Marion, Nubank

ANASTASIA SHVECHKOVA: Okay. I'm joined here by Nathan Marion, head of B2B at Nubank. Thank you very much for joining us today. How is Money20/20 for you?

NATHAN MARION: Thanks for having me. Thanks for the invitation. I think it's really nice. I love the vibe. I used to call myself, like, a payments geek. I really like payments. So just coming to an event like this and just seeing all the industry, the global industry together, is having meaningful discussions. I think it's really, really cool.

ANASTASIA SHVECHKOVA: Great. And you had a session here that was focused on BNPL, correct? What do you think are the biggest fraud challenges merchants usually face with BNPL? Because it gives them a lot of flexibility, but it brings a lot of fraud, and what makes those challenges so tricky to manage.

NATHAN MARION: I think I just wanted to start by separating a little bit of BNPL. In Brazil, we usually say that people have been doing BNPL for a while. So, 'buying now and paying later' has been a reality for the last 50 years in Brazil. And it started by companies like retailers offering credit and giving the capacity of paying installments. But then it migrated to credit cards. So, in credit cards, you can pay in Brazil with installments, and there are interest-free installments. Sometimes you pay for it, your settlement delay for merchants, and everything. And that's where I think a lot of the fraud leaves at the moment. If you think about paying later with cards and Brazil is a little bit of a I mean, it's a market. We play it. It's not for amateurs. I mean, if you're playing there on digital commerce, I think fraud is a big thing.

What we've been building at Nubank, one of the largest issuers in Brazil, 100 million clients, is what we call NuPay, which is the combination of what is best for BNPL and how we can also serve better merchants. It's our own payment method, our own rails. What we've been doing there is trying to address the problem of fraud that exists on cards on BNPL. We give extra credit on NuPay to our clients. So, it's a button on the checkout on NuPay. And essentially, we're also authenticating transactions. So, every transaction comes; this is not on card rails, but it comes back to the app, and it's authenticated. That's the way we're addressing it, just taking fraud out of the equation by authenticating transactions in a very smooth way. It's better than, you don't need to type cards and everything. So, kind of creating this new rails of transaction that address one of those big problems, which is fraud.

ANASTASIA SHVECHKOVA: It makes sense. You know, my Brazilian colleagues always say that Brazilians are too creative for fraud. But you mentioned, of course, that you're building those models. Do you utilize AI right now? And how do you feel AI is influencing fraud in general and in Brazil?

NATHAN MARION: Yeah, we usually say, and this is not new, but I mean, technology is a tool to use for good and for bad. If you think about how the bank was created, several years ago, our mission was to essentially address financial complexity, especially in financial services. Many people lacked access to credit or financial services in general. And this is how we grew in the last years by giving credit to people who have never had it.

But one thing that was very important to us was just building that out, building data. Just understanding behavior. And this is how we build very strong models, very strong credit models that always have used what we currently call AI, but also used machine learning back in the day, everything. So, we're advancing in that sense. We're progressing. We're utilizing new technology to ensure that we can better serve our merchants and users.

That being said, I think AI in fraud specifically is a challenge. Because you can do stuff that has never been able back in the day, especially for identity theft or account takeover and stuff like that. I mean, there's a lot of automation, so we'll need to make sure—not from a Nubank perspective only, but all the industry payments, banking, credit—to uplevel our game also to be able to fight that back and make sure that we have the right systems and ways to understand when there's fraudulent transactions, when there's bad intentions. 

ANASTASIA SHVECHKOVA: It makes a lot of sense. But of course, it's always hard to find a balance between protection and friction for the end user, which you don't want. So how do you think fintechs like Nubank can strike that balance without adding unnecessary friction for merchants or consumers.

NATHAN MARION: One thing I really like about Nubank is that it keeps fighting that complexity, just trying to reach a point to generate some kind of disruption. When we think about NuPay is exactly what we did. It's essentially you click on a button, and you are redirected to Nubank, one of the most trusted brands. You authenticate a transaction; you can pay your account balance. You can pay with credit, card, or installments up to 24, or up to 36. You have access to extra credit. And you don't need to type the card. You don't need to read a QR code, so there's a lot of smoothness, but also adding that level of security, which is authentication. And essentially, you need to create smart friction: what we call using layers of authentication of security when you perceive risk. Not applying to everyone. You need to have very advanced models to do it. But understanding when to do it. It's very important. And trying to rethink the whole process: not only the fraud prevention piece of it, but how it works, the whole payment flow. And what we're trying to address with NewPay is exactly that is really rethinking the flow. As we are giving credit. We can also be the checkout button, and we are in the whole process. So, we can actually enhance the process of payment and give the Nubank experience, also paying outside of our app. So, I think this is one way to see it.

ANASTASIA SHVECHKOVA: With your vast experience with fintechs, fraud, and BNPL. What is one advice you would give to fintechs in 2026?

NATHAN MARION: I think just being just going after the problem you want to solve. Because in the end, if you look again at Nubank's history, it's not that we created an extremely innovativeusers already have product, but really trying to solve the problems we already know users had. Having access to digital service, having access to credit, being able to apply for an account completely digitally, without a physical branch. So, my advice to fintechs would be: identify a problem that users already have and really try hard to solve it. This is a way to be successful in the end.

ANASTASIA SHVECHKOVA: That's very wise. Thank you very much, Nathan. Thank you for joining us today.

NATHAN MARION: Yeah, thank you very much.