Skip to main content
Available on request.

Summary

Return Fraud Detection checks each return against shipment, carrier and return data, and flags the suspicious ones in the Returns Overview with the reason it was flagged. Flagged returns can be routed to manual review, raise an internal notification, and hold the refund until the goods have been inspected.

Who it’s for

Customer service and finance teams who lose money to wardrobing, bracketing, empty-box returns and serial returners, and who cannot spot-check every request by hand.

Key benefits

  • Suspicious returns are flagged, not buried. Each flagged return shows in the Returns Overview with the reason, such as a parcel weight that does not match the outbound shipment, a return filed unusually soon or late, or a shopper who returns a large share of their orders.
  • Auto-approval is overridden for flagged returns. A flagged request goes to the manual review flow even when your rules would otherwise approve it automatically.
  • Your team is alerted. Internal notifications can be triggered for flagged returns so the right person reviews them.
  • The refund waits for the inspection. A flagged return can be held so the warehouse checks the goods before the refund is released.

How it works

A model scores returns on two kinds of signals. The first uses shipment, carrier and return data Perform.AI already holds, such as weight differences and return timing. The second uses consumer behaviour across orders and needs profile data you provide, such as account age, purchase history and return history per customer; the automatic routing of flagged returns to manual review depends on that data being available. For higher-value products, set your refund trigger to fire after the goods are received back in the warehouse. Contact your account team to enable this feature.