Fraudology Podcast with Karisse Hendrick

Welcome back to Fraudology.
I’m sharing the presentation I gave to open the Merchant Fraud Alliance conference in Chicago. It was a “where are we right now” look at the fraud attack landscape, and it grew out of a question I heard from merchants all through the conference. What is actually different about fraud today, and how much of it comes down to AI?
If you work in e-commerce fraud, you have probably felt the shift in your own numbers. Generative AI has changed who can attack, how quickly they can do it, and how well the attacks blend in with good customers. I walk through how AI fraud looks from the ten thousand foot view I get from talking with so many merchants, and I pair it with fraud prevention ideas you can take back to your own team and your own leadership.
What you’ll hear in this episode:
Why fraud has been industrialized, and how AI lowers the cost, skill, and time needed to attack e-commerce at scale
The headline numbers, including fraud pressure up 33% year over year, account takeover up 78%, card testing up 175%, and BOPIS fraud up 65%
How the EMV chip shift in 2015 and the rise of account takeover compare with what AI is doing to the fraud attack landscape today
Why AI-powered account takeover keeps climbing, and how generative AI makes phishing cheap to personalize
How AI card testing attacks differ from old scripts, including how agents sort cards and adapt to your controls
How one campaign can move from identity to checkout to customer support to disputes, and why event-based controls miss it
Why evidence is now cheap to fake, from fake receipts and identity documents to fake damage photos and deepfakes
What a dark web card marketplace demo shows about how easy it is to buy the pieces of an attack
How to fight AI with AI fraud detection using session-level data, telemetry, device ID, and behavioral biometrics

You should listen to this episode if you:
Need to explain fraud increases to leadership and want current numbers to bring with you
Have seen a sudden rise in account takeover or card testing, or heard from your processor about the Visa enumeration program
Deal with refund fraud or chargebacks involving photos, receipts, or item not received claims
Are evaluating fraud prevention tools and wondering why snapshot-based controls keep missing attacks
Want to hear how merchants are using generative AI to spot AI-written scams
Couldn’t attend the Merchant Fraud Alliance and want to hear how the conference opened

Welcome back to Fraudology.
I’ve now had a few days to recover from the inaugural Merchant Fraud Alliance conference in Chicago. I helped create this event. Watching two days of conversation among merchants go from a plan to a room full of people was a real highlight for me.
The Merchant Fraud Alliance 2026 conference was built around a gap I had heard about for years. Merchants wanted a place to compare notes with peers without feeling like the target of a sales pitch. We kept the ratio at roughly three merchants for every solution provider, invited sponsors based on what their customers say about them, and added merchant only deep dives. I explain why we made those choices, then walk through the AI bootcamp, the deep dives, and the main stage sessions. Topics include fraud as a service, app cloning, account takeover trends, executive scorecards, build versus buy decisions, and the PSP gaps merchants keep running into.
This episode is for everyone who couldn’t be there. Two days of sessions won’t fit into one episode. I’m sharing the moments that stuck with me the most.
What you’ll hear in this episode:
Why we built the Merchant Fraud Alliance as a merchant first fraud practitioner event, and how invite only sponsorship shaped the room
How the merchant only deep dives worked, including a fraud center of excellence discussion, a refund fraud ecommerce session for food delivery and retailers, and war room roundtables for retailers and marketplaces
What came out of the AI fraud bootcamp for merchants, from Slack alerts and chargeback research to governance questions and the rising cost of AI tokens
Why a fraud as a service session encouraged attendees to treat fraud threat intelligence from Telegram, Discord, and the dark web as an input to decisions at the time of purchase
What one vendor’s data showed about app clone fraud detection, with chargebacks running roughly four times higher when cloned apps were involved
Ecommerce fraud trends 2026, including a threefold rise in account takeover over two quarters and agent traffic that one vendor found to be nine times riskier than human traffic
Why fraud now moves upstream and downstream of the payment, and the merchant PSP fraud gaps that appear when you rely only on processor tools
Lessons in cross functional fraud collaboration, and how to build a fraud executive scorecard with metrics your executives actually want
A look at the build vs buy fraud stack panel, a non payment fraud discussion covering loyalty fraud and promo abuse, and research on how cyber signals can improve payment fraud decisions

You should listen to this episode if you:
Couldn’t attend and want a practical summary of the sessions and the ideas that came out of them
Are curious what a merchant only fraud conference looks like and whether it’s worth attending next year
Are exploring generative AI in fraud operations and want to hear what other merchants are doing, and what they’re struggling with
Struggle to explain fraud to executives and need a better way to present your data
Are weighing a build or buy decision, or feel limited by the fraud tools your PSP provides
Deal with refund abuse, promo abuse, or loyalty fraud and want to compare notes with peers

Welcome back to Fraudology.
Welcome back to Fraudology. I'm Karisse Hendrick, and this week I'm talking with someone who has become a regular guest on the show. Matt Vega has joined me across six years and several different employers. He's now the Chief Fraud Strategist at Point Predictive, where he works with Frank McKenna on lending fraud, and his move into that world gave me a good reason to ask what looks different from where he sits.
Lending fraud rarely looks like the fraud most merchants picture. Plenty of losses start with an applicant who stretched the truth about income or cleaned up a credit report through repeated disputes, and those cases sit right next to organized attacks. Matt explains how credit washing fraud and income misrepresentation blur the line between abuse and fraud. He also shows how lending consortium fraud data lets a lender see patterns that a single credit report never will.
Then we get to the part I think deserves the most attention. A super prime synthetic identity can now be assembled quickly, and on paper it can outperform a real customer. Whether you work in lending, fintech, or ecommerce, you'll leave with a clearer sense of where your defenses actually hold up.
What you'll hear in this episode:
How lending consortium fraud data works, and why one lender's loss can protect an entire network from the same attack
How credit washing fraud and credit bureau dispute abuse can turn a 580 score into a 790 for a short window, and why lenders sometimes miss it
Why income misrepresentation and bust out fraud lending sit on a spectrum between friendly first party abuse and deliberate attacks
How a super prime synthetic identity can be built in roughly 90 days using an authorized user fraud scheme and buy now pay later fraud
Why a perfect credit profile can be the red flag, and what fraud network intelligence lending teams can see that a credit report cannot
How jailbroken LLM fraud tools, dark web fraud tools, and a dark web identity marketplace make it easier to produce identities, identity document fraud lending, and matching cards
Why AVS CVV fraud limitations and PSP fraud tool limitations show up as high declines, false positives, and chargebacks
How good user behavior mapping makes anomalies easier to spot, and how polymorphic fraud attacks and device farms try to imitate real human behavior
How behavioral biometrics lending fraud controls, friction strategy fraud prevention, and fraud stack vendor evaluation fit together, including why a design partnership fraud tech opportunity can be worth saying yes to

You should listen to this episode if you:
Work in lending, auto finance, or fintech and want a current look at lending fraud beyond the standard synthetic identity playbook
Are responsible for fraud tech stack strategy and want a practical way to think about proven vendors versus newer technology
Are a merchant relying on a PSP fraud tool and wondering why your declines and your chargebacks are both high
Need language to explain friction decisions to executives and growth teams using data they already care about
Want to understand why a fraud consortium lending network matters when attackers are using AI to adapt quickly

Welcome back to Fraudology.
AI is here, and it’s making every kind of fraud we already know cheaper, faster, and harder to catch. We are going to walk through a handful of stories that show exactly how AI enabled fraud is showing up right now. From card testing at scale to one very messy fraud story.
That story is Polymarket, and I spend some real time on it. It’s rare that we get this much visibility into a company’s actual fraud details. A Wall Street Journal investigation lays out the stolen debit cards linked to thousands of new accounts and a fraudulent deposit rate that hit 80%. I also get into AI-generated deepfake delivery orders, a bank coalition’s new warning about agentic commerce fraud, and a warning about AI voice cloning scams.
We aren’t eliminating fraud, we never can, but we can get the information to be as fast or slightly faster than the fraudsters.
What you’ll hear in this episode:
Why AI-enabled BIN attack fraud and card testing are up over 75% according to at least one fraud vendor, and how fraudsters are now optimizing payment authorization the same way legitimate payment teams do
A full breakdown of the Polymarket fraud scandal, including the prediction market fraud risk baked into its business model and the leadership decisions that made things worse
Why Polymarket scrapped its same-source withdrawal rule to speed up payouts, and how that decision opened the door to real money laundering prevention failures
A serious account takeover fraud exploit where a new account created with someone's stolen SSN could inherit their existing balance and linked cards, no password needed
Why compliance staffing fraud programs need people with actual prevention experience, not just investigators brought in after the fact
How growth at all costs fraud risk plays out when a CEO's response to an 80% fraud rate is reportedly "just keep growing and pay a fine"
A new bank coalition report on agentic commerce fraud, and why AI shopping agent chargeback liability still has no real framework under current Visa and Mastercard rules
An AI-generated fake delivery order used to attempt a retail theft at Walmart, and a Georgia police chief's warning
What a real chargeback monitoring program actually looks like from the inside, and why Polymarket landing a fraud vendor like Riskified says a lot about how serious this problem became

You should listen to this episode if you:
Work in payments, card fraud, or retail fraud prevention and want a grounded look at how AI is scaling attacks you already know
Are building or evaluating a chargeback monitoring program and want a real, public example of what happens when fraud rates spiral
Care about prediction market fraud risk or are watching the regulatory and legal fallout at companies like Polymarket
Are responsible for compliance staffing fraud programs and want language for why prevention expertise matters as much as investigative expertise
Are tracking agentic commerce fraud and want to understand the AI shopping agent chargeback liability gap that still hasn't been solved
Want practical, real-world examples of AI-generated fake delivery orders and AI phishing scams to bring back to your own fraud team

I'm joined today by Sudhir Lanka, Associate Director of Fraud Strategy at GrubHub. Sudhir's team doesn't just cover GrubHub anymore. His scope recently expanded to include Wonder, GrubHub's new parent company, and Blue Apron, which means he's thinking about food delivery fraud across three genuinely different business models at once, and I wanted to dig into how that actually changes his approach.
We get into what makes a three-sided marketplace uniquely exposed to fraud, since GrubHub has to protect diners, restaurants, and drivers all at the same time, and a gap in protection on any one side eventually breaks trust for everyone else. Sudhir walks through the primary fraud vectors his team deals with, account takeover, payment fraud, refund abuse, and promo abuse, and gives some of the most specific, real-world detail I've heard on this podcast about how each one actually plays out, down to the exact excuses customers give to get a refund they're not owed.
What you'll hear in this episode:
How Sudhir's career path through JP Morgan Chase, Discover, and GrubHub shaped his approach to fraud strategy at each stage of scale
Why GrubHub, Wonder, and Blue Apron each carry different food delivery fraud risks despite serving the same underlying mission
The three primary fraud vectors GrubHub tracks, account takeover, payment fraud, and refund and promo abuse, and how they show up differently across business lines
A detailed walkthrough of restaurant account takeover, including how a compromised owner's email can lead to a redirected ACH payout and an expensive double payment for GrubHub
Real examples of driver and diner collusion, including self-delivery loops and a surprising exploit tied to minimum wage laws in cities like Seattle and California
Why refund abuse and first party fraud can't be predicted at the time of transaction, and Sudhir's framework for placing controls at the actual point of irreversibility instead
How layered fraud controls work across account creation, checkout, and post-order stages, including multifactor authentication, 3DS authentication, CVV validation, and delivery PIN verification
The difference between soft friction and hard friction, and why Sudhir intentionally reserves harder friction for a very small percentage of customers
Why Sudhir sees fraud strategy as fundamentally pro-growth, not anti-growth, and how protecting trust across the marketplace translates directly into revenue

You should listen to this episode if you:
Work in fraud strategy at a marketplace, food delivery, or platform business balancing multiple user types
Are dealing with refund abuse, promo abuse, or first party fraud and want a real framework for controlling it without over-relying on prediction
Want to understand restaurant account takeover and payment redirection fraud from the platform's side, not just the consumer side
Are building or refining layered fraud controls and want concrete examples of where soft friction versus hard friction actually belongs
Need language to make the case internally that fraud strategy is pro-growth, not a blocker to it