This session was presented at BLUEPRINT 4D 2026.

Product returns are expensive, slow, and manual because someone has to touch and inspect each item before a decision is made. That lag drives up warehouse labor, delays refunds, and leaves your JD Edwards inventory and recovery values out of sync with reality. In this session, we’ll show how JDE EnterpriseOne Orchestrator and Generative AI can pre-classify product condition before the item arrives, using only the customer’s return comments. You’ll see how free-text reasons are converted into clear disposition categories—New/Sealed, Like New, Refurbishable, Damaged, or Non-Resale (hygiene/severe damage)—and then drive an automated workflow in JDE: Orchestrator receives the JSON payload from a third-party returns system, calls GenAI for classification, creates the appropriate purchase order, and uses Receipt Routing to apply the correct status so inventory and refund processes update with zero or minimal manual review. We’ll walk through the orchestration design, the GenAI prompt pattern, error/low-confidence handling, and how to measure business impact in cycle time, labor savings, and recovery value.


This session is eligible for the following CPE Credit: NASBA

Speaker: Kevin Musick, IT Convergence