Fraud doesn't sleep, and neither does the challenge of keeping good customers happy while stopping bad actors in their tracks. In this live episode of Talk Commerce, recorded on the floor of Shoptalk in Las Vegas, host Isaac Morey sat down with Nicole Jass from Signifyd for a candid, fast-moving conversation about what it really takes to protect an e-commerce business today — without driving away the customers you've worked hard to earn.

From machine learning and consortium data models to agentic commerce and instant refunds, Nicole covers the full spectrum of what modern fraud protection looks like — and why it's about so much more than just stopping a bad transaction.

Key Takeaways

Fraud protection should grow revenue, not just cut losses. Signifyd helped Lenovo increase top-line sales by 4–5% by reducing false declines alongside fraud.
The consortium model is a force multiplier. Sharing fraud signals across thousands of merchants means that stopping one bad actor protects everyone on the network simultaneously.
Not all bots are bad anymore. Agentic commerce is changing how merchants need to think about automated traffic, and Signifyd is already preparing for that shift.
Returns are the next fraud frontier. Instant refunds are a loyalty-building tool — but only when merchants can tell the difference between a genuine return and someone shipping a box of rocks.
Scaling fraud teams with people alone won't work. As businesses grow, "Katie reviewing transactions" isn't a sustainable strategy. Machine learning fills the gap efficiently.
The customer experience IS the product. Every decline, every friction point, every delayed refund is a moment of truth that either builds or breaks brand loyalty.