Trust by Design: Rewriting Authentication in the Age of GenAI Fraud

0:00:09.0 Kailey Raymond: Welcome to Builders Wanted, the podcast for people shaping what's next in customer connection and brand growth. Today we're talking about digital trust, why it's breaking, how it's being rebuilt, and what leaders can do about it. Fraud has accelerated with Gen AI, deepfakes, and scams that hit not just revenue, but real people. Our guest is Filip Verley, CIO of Liminal, who advises on identity and fraud prevention strategies worldwide. We'll explore where the cracks are showing, how the best teams are responding and why unified platforms are reshaping the future of authentication. Let's get into it.

0:00:45.7 Producer: This podcast is brought to you by Twilio, the customer engagement platform that helps businesses turn real-time data into seamless, personalized experiences. Engage customers on their terms across SMS, voice, email, WhatsApp, and more. Power every interaction with AI so conversations feel natural, not robotic. Adapt in real time, delivering the right message on the right channel exactly when it matters. That's the power of Twilio. More than 320,000 businesses from startups to Fortune 500s trust Twilio to transform customer signals into conversations, connections and real revenue. Reimagine the way you engage with your customers. Learn more at twilio.com.

0:01:34.4 Kailey Raymond: Filip, thank you so much for joining us on the show today. I'm excited to learn from you.

0:01:38.9 Filip Verley: Awesome. Thank you for having me. Really appreciate it.

0:01:41.0 Kailey Raymond: This will be really great. And you know, I'm not a fraud expert, so you're going to have to help define some of these terms we're going to talk about. But just as a brief introduction to the show, obviously, Filip, you're the CIO of Liminal. You're on the front lines of fraud evolution, which is really what we're going to be talking about today. Just to kind of kick us off, what are you hearing from companies that's keeping them up at night? 

0:02:05.4 Filip Verley: Yeah, absolutely. And even though you maybe you're not a fraud expert, Kailey, but how many spam messages are you getting daily? 

0:02:12.2 Kailey Raymond: So, so many.

0:02:13.3 Filip Verley: Right. And then not even talk about the calls, right? If you were to answer one of these spam calls, would you actually know that that person is a real person that works for JPMorgan Chase or is out of some farm operation in Southeast Asia, where they have now deepfakes that can really mimic and really look and act and sound like whatever individual you like? And so that's really what we're seeing a lot, is this Gen AI. Gen AI is really getting people very nervous because it's extremely hard now to figure out what's real, what's not real. And so that's again, is like fraud has evolved so quickly and our defenses haven't really caught up with them. And so we're seeing, you know, in the report, we surveyed a bunch of practitioners, 71% of companies say they've lost customers due to false declines or unnecessary friction. Because what do you do when you see a lot of fraud? You tighten, you tighten it up. And that's again where you see that bad experience for good customers and you're not catching any of the fraud. So again, long-winded way of saying Gen AI is changing the game. And really, practitioners are very concerned about how they can actually address these concerns. And there's no real good answers that, and they're struggling. They're basically struggling.

0:03:33.6 Kailey Raymond: One hundred percent. I heard about this one of one of my friends, her parents are Vietnamese and so they're immigrants to this country. They had a fraud call that was sent over to them that sounded like their daughter, my friend. And it's just a really scary world, you're right, with Gen AI out there with the capabilities today. And what you're talking about is it's hard to keep up. There's some of this loss of trust or loyalty, which is really the thing that's on the line right now. It's what we're all playing for as long term customers, but we don't want to lose their trust. And you just quoted a stat from your team's seminal report. It's the industry benchmark on fraud and identity. And so the report also says that 78% of businesses experience repeat account takeover, which is when an attacker takes control of a user's account. Maybe they're going to steal money or data or reputation or whatever it is that points to the loyalty kind of conversation that we're talking to right now. So it's not just fraud, it's that loss of trust. So why is this such a hard problem to solve? 

0:04:37.5 Filip Verley: Yeah. If I had the answers, I would, you know, you and I could start a business.

0:04:42.6 Kailey Raymond: Help us solve it today, Filip.

0:04:43.1 Filip Verley: Yeah. Let's solve it today on this podcast. We'll just knock it out. It's really hard because the threats are consistent, persistent, and they're just very adaptive. I worked at Airbnb. Airbnb, all they sell is trust. They don't have inventory, they don't own houses, they own hotels. They just sell trust. Basically, can I trust Kailey to come to my house, hang out with me in my house, or vice versa, or Kailey has advertised that She's a helicopter pilot. And one of the experiences in Airbnb is to fly across San Francisco in a helicopter. If that was on Craigslist, I probably wouldn't do it. But Airbnb has this brand reputation that says, hey, we connect people. And so that's all they do is sell trust. And so at those times, we're like, oh, we have an account takeover problem. Because hosts make a lot of money on the platform, and somebody can take over that account, funnel the money out, but they can also take over these accounts and publish fake listings, which then people book. And then all of a sudden, your money's gone. And so it's a massive problem. And so, oh, account takeover is a problem.

0:05:52.8 Filip Verley: Let's push the lever and make it really hard to take accounts over. Then account openings becomes a gap, because now you're not focused on account openings, and then all of a sudden you pull that lever. So it's all about, how do you balance this experience with the risks that you're seeing? And still it's all about that good customer. The good customer is the one that gets the impact most, right? It's always that customer. And so this ATO piece is really hard because you're tightening the controls. But are you tightening the right controls? And are you just over, you know, are you... Man, I can't even think of the word. Are you over... What is the word? 

0:06:30.0 Kailey Raymond: Over indexing.

0:06:31.4 Filip Verley: Over indexing. Thank you so much. Yes, over indexing on that one thing. And so the other problem is you have a lot of teams in your organization. You have a fraud team, you have a customer experience team, you have a growth team. Again, Airbnb, we had a trust and safety team. We had a host team, a guest team. If you have teams, you create silos. If you create silos now, all of a sudden, data, information does not flow. And even, I'm sure at Twilio, even at Liminal, we're very, you know, we're 75 employees. But you're seeing the silos seen created by, Oh, we did research on that. Oh, I didn't know we did research on that. Why did I not know that? I'm in the company, why shouldn't I know that? So silo teams create these silos of information data. And so now all of a sudden, you are fighting a very organized organization because they are organizations and not someone in their basement keyboarding away. This is an organization that have offices where people come in nine to five to commit fraud. Organized, focused, whereas we're like, hey, we have these cool teams and they're not talking to each other. So we're actually giving them an advantage by being disorganized.

0:07:46.1 Kailey Raymond: You're bringing up a lot of themes right now. So you're talking a little bit about silos, you're talking a little bit about this tension between tightening the controls and the reality of growth of a business. I think these are tensions that we're going to continue to explore throughout this conversation. So a good kind of setup for us for some questions coming up. Going back to the report for just like one second, you know, we talked about this like account takeover. You've mentioned a couple of stats. I'm wondering, what's the one that surprised you the most out of this report that you just published? 

0:08:17.0 Filip Verley: There was a bunch of really good ones, but I'd say that the big shocker was 87% of practitioners don't feel prepared for transaction-related fraud. And it goes back into where we started with Gen AI, deepfakes, folks just do not feel prepared with what they have in place today. And then the other one I would say is that, and it goes back to what we just talked about is like 65% of them say that their trust insights are trapped in individual systems or teams. They are saying, yes, we should have a consolidated approach. No, we have multiple tools and we have multiple teams. So having that understanding, there's a high awareness of the issue, but very much a lack of preparedness in trying to fight or combat that. And so that was one that I'd say that's a very massive takeaway.

0:09:09.5 Filip Verley: I know you asked for once and I'm giving you a third one since I'm a bad listener. But 92% of practitioners want multi signals. The report is all about convergence of authentication and fraud. Having signals is really the key because if you're looking at signals, you're looking at patterns. If you're looking at patterns now, you might be able to detect some of this deepfakes, some of this fraud because they can pass one, two rules. It's once you establish a pattern that it could fail those patterns. So again, gave you three stats, but again those all kind of match and connect together. And it was super eye opening to see that folks know they have a problem, they just don't know how to tackle it.

0:09:52.9 Kailey Raymond: Yeah. I mean, it's like a little bit of a scary world. But luckily I do think that there are playbooks that are kind of emerging right now. One of the things that you're talking about that strikes me is data unification and making sure that you have that really kind of tightened up is something that kind of comes top of mind when it comes to part of the playbook. One of the stats that struck me was that 71% of buyers also worry that they can't stop social engineering today. It's just something, it's the sphere that we've kind of been talking about. But I want to talk about the playbook. How are attackers changing their playbook and then how are companies responding to this? You talked about Gen AI a little bit. Let's talk a little bit more about it.

0:10:34.2 Filip Verley: How are they changing their playbook is because they can, because they have the tools now. These AI platforms, these large language models have democratized fraud. You and I cannot become experts at fraud. I've been in the fraud space forever. I've never been able to pull off a good fraud trick. You know, I always catch them. I try to catch them, but it's like super easy, right? One of the things that is not in this report, but it's something we've done studies on as well, is maybe five, 10 years ago, it was the Nigerian prince scam or the romance scam, right? You got all those emails and like, oh, if you can send me some money, I'm trying to move this money out of Nigeria because I'm being persecuted. I have $20 million in gold and if you just give me $2,000, I will cut you in. Littered with grammar. You were just like, what is this mess? Because of gen AI, because of the LLMs, the grammar in those types of emails have now jumped by 30, 40%. So the improvement in grammar, you cannot detect it. And now just take a step back and you can look and say, okay, let me go into ChatGPT right now.

0:11:44.7 Filip Verley: If I go into GPT right now, say, hey, I am a Bank of America customer service representative. I am going to be reaching out via email to my customers to get them to change their password. I am going to add the link in there and I need it to sound exactly like the tone of the company and that it feels and sounds and looks like a very safe message coming from Bank of America. Sounds great. Let me help you with that. LLMs never said they won't help. They always want to help you. And so now all of a sudden you can not just send an email, you can impersonate. And so that's again where these playbooks have been changed because they have now these tools at the tip of their fingers. They don't just have to create fake information. They can impersonate you and I, they can just take my face, take my voice, and all of a sudden, just like your example, make a call that sounds exactly like me. And I'm like, well, I'm not in jail in Algeria and you don't need to send $10,000. But yes, it's just like that. Democratization of these tools have given them the opportunity to really take advantage of where your gaps are, which comes back to that point we just talked about.

0:13:00.0 Filip Verley: These siloed teams, siloed data, siloed information is absolutely playing in their hand. And that's where, one way we've seen where you can kind of battle that or defeat that is can you have a company wide goal of trust or integrity where every team, marketing growth, customer service, fraud, cyber, you name it, they all have to tie into that one goal. And if you can do that, then all of a sudden now we're starting to take the same language, we have to share information because we're all tied to that same objective. And so that's one way that you can kind of as an organization start looking at that. Because then you're opening up signals for your fraud team to the marketing team, from the marketing team to the customer success team, et cetera. And now you've got the flywheel going. Now you're using all the information you can to again combat that very organized attack.

0:13:56.5 Kailey Raymond: One hundred percent. I like what you're talking about, which is true across most organizations. Being able to come together and unify across teams is inherently important to being able to kind of drive direction. And we're talking about some of this unification with data. We're also talking about it with platforms. And we know that a lot of companies have two, three, four different places where they actually might be keeping some of their data, where they might be kind of dispensing across different fraud solutions, these little kind of point solutions. So when you think about that, this juggling act that people are doing when it comes to authentication and fraud, where do you recommend they start when it comes to unification? 

0:14:41.0 Filip Verley: What we're seeing a lot is like, you don't have to rip and replace, right? You don't have to be like, oh, I got to start from scratch. We're not seeing that. What you need them to look for is, how can you connect the dots across your system? Where are you making key decisions, are those key decisions at onboarding, are they at the login place? Are they at the checkout? Account changes? And ask yourself, are those systems talking to each other, or are they very separate, unique places where, again, the data is not talking to each other, signals are not getting shared. And so that's where you're going to try to connect those device signals, the behavioral signals, identity signals, connect them into one system.

0:15:22.3 Filip Verley: So you want to rip and replace? Great. But again, that's pretty risky. Try to understand, can you actually connect some of these disconnected tools as a starting point? Once you've done that, then you can start building out and saying, okay, how do I now work and optimize good customer experience? Do I want to optimize for that? Do I want to optimize for fraud? And I think again, you decide as an organization, you decide, what is the balance you want to strike? Do you Want to do it 50 /50? Do you want to do 60/40? You're in hypergrowth mode. You don't care about fraud. You're 100% customer experience. It's up to you. But if your systems don't talk to each other and they're disconnected, you're going to run out of juice and you're going to lose all your money.

0:16:07.3 Kailey Raymond: Totally. I mean, one of the things that you're kind of mentioning in here is if you're connecting these systems, you're getting a lot of signals, a lot of different types of signals that are helping you to identify risk in some kind of different ways. I'm wondering, with all of this data, there's probably a lot of noise in there too. Like, not all data is necessary. Some data is probably more helpful or interesting for certain uses than others. So if you had to pick one, one signal that really kind of changes the game, which would it be and why? 

0:16:41.3 Filip Verley: The one that's really hard to spoof is intent, behavioral intent. If you can get a credential, you can clone a device, you can get all of those things. But how someone is moving, clicking, how they respond to friction, a lot of behavioral analysis and intent is really hard to do at scale. Can you do one at a time? Yes, you could, but that's not what pays off for these organized groups. They want to do scale. And so if you're thinking about, like, Filip goes into his bank account, he always uses his password manager to autofill. If I all of a sudden started typing my username and password. A little bit odd, right? That'd be odd. That would raise a flag for the bank. Next thing I do, I typically go and check my balances and I scroll to the bottom of the page and then I go, okay, I need to pay a bill versus I immediately go to pay bill and do wire transfer. Those are behaviors that is not typical intent of what Filip does on that site. And so if you put that together with how I'm navigating, that's where you have really good, strong signals that again, can they be broken? 100%, absolutely, but not at scale.

0:17:58.9 Kailey Raymond: Never thought that Bank of America knows that I have really similar patterns. Of course they do.

0:18:04.7 Filip Verley: Of course they do.

0:18:05.8 Kailey Raymond: And of course they're taking a look at that. That's a really interesting insight. I also think that one of the interesting ones that you kind of pointed out, and I think there's a lot of tension around, is this idea of, users don't like friction. Behavioral intent. You're trying to basically find the pockets of like, where can we reduce friction? How can we increase growth? Right? But at the same time, sometimes you to increase friction to make sure that you are preventing fraud. So what is this tension that we're talking about? How do leading teams decide where you draw the line between speed and safety? 

0:18:43.1 Filip Verley: The best teams or organizations don't think in an either or. It's the balance. They always are able to balance. And they design these systems to adapt to what they need. As I mentioned earlier, with this lever pulling, account openings, account takeovers, they know the exact formula of how much they need to tweak a certain threshold or a certain friction to balance against the experience, to balance against the fraud. And so I would say there, it's not just about reducing the fraud. It's making sure that users are protected without slowing them down. And so smart friction, smart friction, what does that really mean? It's basically, again, taking another previous role I had. I worked on age verification. And age verification at Google, specifically for YouTube and YouTube's like, hey, don't mess away with my watch time. If you mess with watch time, just hand in your badge. Because watch time equals money and it's growth, et cetera. And so the big thing for us was we needed to balance the regulatory angle, right? We're talking fraud here. But this is, it's very similar.

0:19:49.5 Filip Verley: The regulatory compliance. We had to be compliant with the regulation, but we have to balance it with the good experience of customers. And so the way we did that was like, okay, well, if we put it right in front of a video, how many YouTube videos, you watch a day, five, 10 videos and next thing you know, video 11, I'm like, oh, sorry, you need to confirm your age. Can you please provide us a government issued ID or your selfie? That is a terrible context. The context is not there. The friction is heavy handed. So you got to look for what is the user's context that they're in, what is their awareness of what they're doing. And so that's how you can balance the friction with the good experience.

0:20:31.7 Filip Verley: Take an example again of a bank. The friction should be at login, making sure you are the right person and then let you go and say, oh, I want to create a new account, open a new account. Why would you put more friction there? But maybe if you're doing that wire for 5000 that you've never done, that's where friction's appropriate. Because now as a customer, if it's a good wire, I'm the one actually doing it. I'm like, okay, they kind of have my security in mind. They're making sure it is me before I send this 5000 out of my account and I'll never see it again. So it's that balance. Can you balance where the friction and the sense of security makes sense, where the customer is in that journey? 

0:21:13.1 Kailey Raymond: That's really interesting. I also really like that you brought up age controls and things like that because that's a really hard one to solve for because there's such like, these companies obviously are incentivized to get people to view, to your point as many hours as they possibly can and get people onboarded as young as humanly possible. One of the elegant solutions that I think is now a new proposed law in California is like, hey, what if we could just get the parents to tell us what age the person is? Wow, what a concept. Which is like friction in the right place up front, right? It's like you can't do it until somebody that we trust to tell us your age is going to tell you that. Right? So I think that's really interesting.

0:21:54.4 Producer: Great customer experiences aren't magic, they're built. And Twilio is the platform that helps you build them. Every customer action, browsing your site, opening an email, reaching out to support, triggers instant AI-powered personalized engagement across SMS, voice, email, chat and more. No delays, no guesswork, just the right message at the right moment. From automated messaging to seamless authentication to hyper personalized customer journeys, Twilio's customer engagement platform powers millions of interactions daily, helping businesses drive loyalty, optimize marketing spend and create experiences that people remember. Twilio is the ultimate toolbox for customer engagement. Ready to build experiences that matter? Visit twilio.com.

0:22:45.8 Kailey Raymond: I want to get back to this idea you said kind of upfront with that question is like, we can't really think about it as a trade off. The best companies need to think about both things at the same time. And so I'm wondering, if you have to define kind of the basics of this for folks, what are the layers of defense that every company needs from the very beginning of login to recovery to making sure that we're keeping sessions safe. What are the full layers of defense that folks need? 

0:23:13.5 Filip Verley: Yeah. I would say that there's definitely four layers that we can always point to. But again, I would say that even in your own journey, in your own company and the products you are offering to customers, they could vary a little bit. But the four ones that we always look at are making sure that at login that you have strong credential and device checks. Right? Making sure that the information is correct. But then also, have you seen this device before or is this a completely new login? It's no longer from Filip's iMac on a Chrome browser. All of a sudden he's on a Chromebook and it's like odd, right? So making sure that you have those at login for sure. The second part is monitoring the user's intent, understanding that behavioral thing we've talked about a little bit already, but behavioral intent, behavioral monitoring is really key. Then real time risk detection across sessions, right? Can you actually detect what is happening? Is this an anomaly? Can you friction them at that point? Do you feel like you want to friction them? Or again you're like, I'm leaning towards a good experience, I'm going to let this go. But real time monitoring is key.

0:24:25.0 Filip Verley: If you're just doing it at the moment of login and you're like, oh, my job is done, that's not good. You have to continuously monitor that journey of that customer. And then the fourth piece is, let's say you're pretty sure that Filip is a fraudster and you knock him out of his account. Like, I get logged out of my account, how can I recover my account? The account recovery piece. Because again, fraudsters are not going to come back to recover. It's the good customers that come back and say, hey, what happened? Like, I need to get into my account, it's Friday and I need to go to happy hour. I need some money. But imagine I'm knocked out of my wallet on my phone. Now all of a sudden my digital life is stuck. You have to have an ability to recover. And so recovery of account is that fourth angle. You can't have the first three and not have that last one because otherwise again, you are going to lose a lot of good customers. And again, as we mentioned before, 78% of companies see repeat account takeovers right from the same actors. So it is really important for you to have a good story around the recovery and not just about the account takeover defense.

0:25:36.5 Kailey Raymond: I love the examples that you're giving. You're right, I'm not a fraud expert, but I felt literally almost every single one of these things at one point in time. So it's really easy to put yourselves in the shoes right here and understand all of these examples. So I very much appreciate it. I also happen to know that budgets, which makes a lot of sense based off of this conversation, but budgets are going up in this area. So where are companies investing to make the biggest impact? 

0:26:05.2 Filip Verley: Yeah. Budgets are going up. We are seeing budgets going up in other research we've done. It's a big theme for '26, is that fraud budgets, AML budgets are all going up. The investment, the shift is towards convergence. How do we converge some of these tools that we have to make better, more contextual decisions than we do today? Even though more than 50% teams still use three or more tools to manage fraud and authentication, they're all seeing the bigger picture. They have to get to this reduction of complexity by having multiple tools and trying to invest in platforms that can unify a risk score, authentication, and identity resolution. How can you combine that into one? And if you're not there again we earlier mentioned, this is like, at least start connecting the dots and figure out are you with the right vendors? Are you getting the right signals or do you need to go find someone that could do this risk scoring, authentication, and this identity resolution piece? So that's where we're seeing a lot of money being poured into, is like figuring out what is my right solution to get the most signals to make the best decisions for the business? 

0:27:19.5 Kailey Raymond: So, when you're moving from these patchwork tools into this unified identity platform, what are the choices that make the biggest difference for folks? 

0:27:32.0 Filip Verley: The biggest difference comes from being able to have that balanced between the security and the customer experience. How are you able to get that balance going is by having that investment. And one of the things we're seeing is 65% of teams say their trust insights are stuck in individual systems. So they have data, they have the information. They're just unable to really harness it and get use out of it. And so that's again where you're having these signals but they're trapped. They're in these disconnected pieces. So that's again where we're seeing a lot of this focus.

0:28:13.3 Kailey Raymond: I want to go back to signals for a minute. I think it's just so interesting, some of these insights that you're sharing. Passive signals is where I want to head, which is, maybe device reputation or we talked a little bit about user behavior. What are some of the things that you're seeing that are helping spot abuse early on without slowing down these real customers? 

0:28:33.8 Filip Verley: It depends on where we go in the journey. So if we're starting, let's say you are a new customer and you are on an e-commerce site, that is very different from, hey, you're a returning customer for a marketplace, like let's say an eBay or something. And so it's all about the context of the journey and the moment that you are looking at the customer. And so these passive signals are really all about what is your intent? Basically, this is what a good user looks like and this is their intent. Do you match that or are you an outlier? And it doesn't make a lot of sense. So let me give you an example. Let's say you are a marketer for eBay and you are, you know, they're big into sneakers. I just find... Auto parts too, by the way. I'm like, wow, that's interesting markets.

0:29:22.8 Kailey Raymond: Well, that's like, it's like hobbyists. That actually makes a lot of sense to me.

0:29:26.1 Filip Verley: Yeah. And so let's say you are like, you know what, we want to boost our auto parts business on eBay. So you're going to publish, you're going to do marketing on Instagram for, do you call them gearheads? I feel like they call them gearheads. But you're going to market to folks on Instagram, TikTok, whatever your channel is, to that community, that folks that are really excited about that. So when you come onto ebay from one of those ads, that is your first signal. That is your first passive signal. Hey, they clicked on an ad from us. And so now I'm like, okay, cool, maybe I won't friction them just yet because there's a high intent. They are very committed. They are our ideal customer. So why would I do friction for them? And so now you go through the motions and you're buying, I don't know why I used cars because I have no idea how to fix a car.

0:30:20.7 Filip Verley: Let's say you're buying a new, something really expensive on a car and so you're buying that and all of a sudden you're like, okay, good intent. There was nothing funky about this customer that was coming in. Now the friction could be right before they're ready to pay. But let's say this other customer comes in and it's on a iPhone and it's like www.ebay.com and they're going to the car parts and they want to buy the carburetor. That's a very different way. And it's again, it shows that passive signals, all of a sudden they start immediately mattering from beginning. And so that's when you're looking at like an e-commerce or a purchase. If you're looking at logins, you have a ton more information because you've already created an account. You've come back before potentially. And now you have device signals. You have their activity signals, like their pattern of how they use the phone. How do they hold it? Do they hold it straight? Do they hold it down? And so you get a ton of those signals for login.

0:31:23.9 Kailey Raymond: What? Do they hold it straight or do they hold it down? 

0:31:28.3 Filip Verley: The strength of your typing, like how hard do you type on your keyboard or on your phone? Are you scrolling up? Are you using one finger, two fingers? 

0:31:39.3 Kailey Raymond: You are blowing my mind.

0:31:42.8 Filip Verley: Ton of signals that are being used. And so again, that's, again, that was a super long answer. But it depends on where in the journey, what type of company you are. Is it a marketplace, is it e-commerce, is it a bank? And that's where you are now starting to put those signals together and understanding, yes, that is Kailey or something's off. Let me do a multi-factor authentication, Send an SMS or send a QR code to take a selfie. All of these little frictions to make sure that it is you or not. And that's again comes back to that balance. How do you feel about the experience versus the fraud protection angle? 

0:32:25.5 Kailey Raymond: These examples that you're giving are such interesting and granular sources of data and they're so mapped to a really granular customer experience. Like, you know the map of this customer's journey to like every single moment that this person interacts with you, that's really, really advanced. And so that's just something that keeps coming up for me is like, these are incredibly advanced things for folks to get into. Like, what's the simple place to start? You know, like, whoa, this can feel so, so overwhelming. What's the crawl version of this run that we're talking about? 

0:33:09.2 Filip Verley: Yeah. Yeah. Read the report, maybe, I don't know, shameless plug. But I would say the crawl version of this is like, connect with people in the space. Just like we've been talking about communication and sharing signals, sharing data, talk to folks that know what they're talking about, right? That have been in the space. You know, I'm sure Twilio, Anurag, Anurag on your team, amazing experience. Like, folks should pick his brain. Reach out to me, pick my brain. What do I do? I pick other folks' brains, right? And that's how you start, is like just starting to get some knowledge. And the one thing, when I was buying solutions, I bought a ton of solutions. When I was again at Google, Airbnb, all of those places, we didn't just build, we also bought. I don't look for vendors, I look for partners. I look for folks that really, truly understand my pain, that can help me solve that pain. And so a lot of people, oh, I don't want to talk to vendors. You know, they always want to sell me stuff. You know what they do? They sell to everybody. They know the industry, they know exactly what's going on with your competition.

0:34:15.4 Filip Verley: They won't tell you everything, but they'll give you really good insights and nuggets of information. So I think that's where you start. And then you look for, as we mentioned, is can you find these platforms that do the risk signals, the authentication, the authentic resolution. Start maybe something with out of the box. And once you feel comfortable with that, now all of a sudden, maybe you start your own data analysis team internally. And those folks are smart. They're all really smart. They'll start finding signals that are really just for your business that nobody else has. So I think it's a journey that educate. There's no stupid questions, right? The one thing I always heard from someone, I was like, yeah, it makes total sense. Like, hey, we're all competing for good customers. We should all collaborate to keep the bad customers out, aka the fraud. And so we should. We don't do it enough.

0:35:06.7 Kailey Raymond: I like it. That's practical. You're right. It's simple. Talk to people, learn, educate yourself. I am going to end us with a couple of quick hit, fast little questions before we wrap up. So you ready? You ready for this? 

0:35:22.6 Filip Verley: I hope so.

0:35:23.3 Kailey Raymond: Lightning round.

0:35:24.4 Filip Verley: Lightning round. Let's go. You've seen from my answers, there's nothing lightning about this.

0:35:28.5 Kailey Raymond: I love it. Okay. What's one emerging threat that you're watching closely? 

0:35:34.4 Filip Verley: Just the overall sophistication of these models. Have you tried Sora? 

0:35:40.6 Kailey Raymond: I have and it is pretty gnarly.

0:35:44.3 Filip Verley: It's awesome and amazing. But then put your fraud hat on and now think of all you can do.

0:35:49.1 Kailey Raymond: You're deepfake, you're, yeah, yeah, yeah, totally. Actually, while we're on the topic of Gen AI, I have a tangential question off the back of that. How do agentic browsers like Perplexity fit into this idea around fraud when you can get agents to just do things on your behalf? Well, how are you thinking about that? 

0:36:09.4 Filip Verley: Yeah. There's a lot there because let's say 12 months ago it was bot equals bad. You could just knock bots off your platform because you didn't want bots, you know, scraping. You didn't want them to come on. And now it's like, oh crap, bots could be good because Filip and Kailey could have their own bot to do shopping, right? They are monitoring prices and all of this stuff. It's a very hard problem because there's a lot of companies in this space that were a terrible way to say it. Basically saying a bot is bad, then I can knock that out for you. Now they're all having to re-figure out and re-configure themselves to understand bot authentication, the authorization, the verification of the bot, tying it to a human. And right now what we're seeing is a lot of websites are like saying, hey look, we're not ready yet. We're going to just still reject bots until we have a solution that we can actually have a communication with the bots that understands what is their intent? What are they authorized to do? 

0:37:10.2 Kailey Raymond: I keep thinking about this idea that there's going to be two internets and that there's going to be an internet for bots, there's going to be an internet for humans. I'm just like, it's just so progressing so quickly. It's so interesting.

0:37:20.1 Filip Verley: You don't need UI. You don't need a UI for bots.

0:37:22.7 Kailey Raymond: Exactly. It's just APIs. It's just. Yeah, totally. 100%. Okay. Don't worry. I'll get back to my fast questions. The next one. What's a company or initiative that you admire for pushing identity fraud? 

0:37:36.1 Filip Verley: I think some of the organizations, like the Fido Alliance, that are really trying to standardize things, and again, you know, they started with passkeys, but you can see the bigger identity picture evolving there. So I think it's the... Because if you don't have standards, you just have the Wild West. And so, again, with all of this AI and the bots and all those things, you have to figure out what is that standard layer that we all adhere to, and start from there.

0:38:01.9 Kailey Raymond: Totally. What's one thing leaders should start and stop doing? 

0:38:07.1 Filip Verley: Start doing, I would say take an inventory of what you have. I would almost question everybody that's listening. Do you actually know who your partners are? Do you know what teams are actually working on your customer journey? That is one. Start understanding your landscape. What are you actually working with? And I think stop, maybe stop buying into the hype or stop getting so concerned about AI and actually do something about it.

0:38:36.7 Kailey Raymond: I like it. And that's true of using it, too. It's like, not only preventing the fraud, but, like, stop the AI hype of, like, we gotta do this and just, like, get some simple use cases and ship them.

0:38:48.5 Filip Verley: That's right. Totally.

0:38:50.4 Kailey Raymond: Okay. Last question. What's the best advice you've received on balancing security and growth? 

0:38:57.7 Filip Verley: I think that some of the best advice I got was again, during the Airbnb days where it was, everything was about the experience, and it was really like, hey, Filip, put yourself into customer shoes. They have one or two vacations a year. They have a lot of choices to pick, platform, destination, you name it. You need to make sure that when they come to our site, they're able to make that happen. And I guess the nugget of that is, like, put yourself in the customer's shoes. Which, again, sounds really silly, but it kind of applies.

0:39:37.4 Kailey Raymond: It's usually the advice that it boils down to. It makes perfect sense. It's always there for a reason. Filip, this was a fun talk about fraud.

0:39:47.0 Filip Verley: Awesome. Awesome. I appreciate it. I'm glad it was fun.

0:39:49.9 Kailey Raymond: Thank you for being here. I really appreciate the time.

0:39:52.0 Filip Verley: Pleasure is all mine. Thank you so much for having me.