Your Fraud Stack Wasn't Built for AI Agents. Here's How to Catch Up | Presented by SEON
|AI agents can now browse your storefront, fill out forms, initiate checkouts, and manage subscriptions without a human ever touching a keyboard. That's a legitimate opportunity for commerce — and a new attack surface most fraud stacks weren't designed to handle.
This session breaks down how device intelligence detects non-human sessions in real time: identifying known AI agents, spotting the virtual device farms behind large-scale abuse, and reading behavioral signals that reveal automation even when it's designed to look human. You'll leave with a practical framework for stress-testing your own defenses.
Attendees will learn to:
1. Explain the emerging fraud risk posed by AI agents in ecommerce environments, including how they browse, fill forms, initiate checkouts and manage subscriptions autonomously.
2. Describe how device intelligence detects non-human sessions in real time, including identification of known AI agents by name and virtual device farms used in large-scale abuse.
3. Identify the behavioral signals that indicate automation, even when it is engineered to mimic human activity.
4. Understand how AI agent detections integrate with scoring engines to trigger custom risk rules without introducing unnecessary friction for legitimate customers.
5. Apply a practical framework for stress-testing their organization’s fraud defenses against AI agent-driven attacks.
6. Distinguish between legitimate AI-assisted commerce and malicious automated abuse, enabling more precise and effective fraud prevention strategies.
