Why AI Can't Fake Where You Are: The Future of Fraud Prevention

In this episode, Arthur Bedel sits down with André, Co-Founder & CEO of Incognia, to explore how location-based identity is reshaping fraud prevention, and why the future of security lies in signals that AI can't fake.

André shares his journey from a family of computer scientists in Brazil, to a college research project on indoor location technology, to arriving in the U.S. with two carry-on bags right as the pandemic wiped out 95% of his company's revenue overnight, and the bold pivot that turned Incognia into a global fraud-fighting platform protecting over a billion users.

Together, they discuss:

• Why the identity signals we rely on today (facial recognition, document verification, device IDs) are increasingly being spoofed by AI, and why physical location is the one signal that's hardest to fake

• What APP (Authorized Push Payment) fraud really is, why bank impersonation scams are so effective, and why consumer education campaigns keep failing

• How Incognia detects mule accounts by analyzing location behavior, including a real case of 200 devices in a single apartment accessing 4,500 accounts

• Why 70% of fraud devices are factory-reset every week, and how location data links a "clean" device right back to the fraudster
• Why detection without a disruption mechanism is just reporting, and what smart, low-friction disruption looks like on the UX side

• How AI is supercharging fraud on every front: near-perfect deepfakes, hyper-personalized phishing, and fraudsters who share tools freely because they don't compete with each other

• The three layers of the agentic fraud problem, binding an agent to a real person, "Know Your Agent," and verifying intent, and where blockchain fits in

• The one thing André wishes every payment risk executive understood about APP fraud, and why the identity stack doesn't need reinventing, just the right layers, in the right order.

Originally produced by Connecting the Dots in Fintech