How the fraud agent works
Fraud agent on every return
Redo's fraud agent runs alongside the return review agent and evaluates every return for fraud signals, including serial return behavior, suspicious patterns, and known fraud indicators.
HIGH, MEDIUM, LOW risk scoring
Every return is scored and tagged with a risk level before it reaches your team. Merchants can filter their return queue by risk score and prioritize review accordingly.
Serial returner identification
Redo detects customers with repeated suspicious return behavior and flags them in the system. Custom policies can be applied to flagged customers automatically.
What merchants control
Auto-approve, auto-flag, auto-deny rules
Merchants set rules that tell Redo what to do when the AI makes a recommendation. Route low-risk returns to auto-approval, send high-risk returns to manual review, or auto-deny claims that hit your defined thresholds.
Included in the base tier
Fraud detection is available to all Redo merchants, not locked behind a premium plan. Loop Returns restricts fraud detection to Advanced+ subscribers.
Feedback loop to improve accuracy
Merchant actions on flagged returns feed back into the model. When you approve or deny a flagged return, Redo learns from that decision to improve future scoring accuracy.
