All use cases

Returns & Refunds Desk

RMA flow with policy checks built in.

Workflows
Prompt
Build a returns and refunds desk for an e-commerce store: 1. Orders in SQLite (seed 20 with items, dates, and statuses) 2. Return request flow: order lookup, item selection, reason codes, photo note field 3. Policy engine: auto-approve within 30 days and unopened, flag for review otherwise — policy rules visible and editable as data 4. RMA queue: pending review, approved (with generated RMA number and return instructions), received, refunded, rejected with reason 5. Metrics strip: return rate, top return reasons, refund total this month 6. Printable RMA slip per approved return Seed requests in every state so the queue demos itself.
Use this

Why this exists

Returns are where e-commerce margins quietly die, not in the refunds themselves, but in the handling: the email threads, the "did we approve this?" archaeology, the policy applied differently by whoever answers. Dedicated returns platforms exist, priced as a percentage of your problem. This blueprint builds the desk itself: your policy as code, your queue, your data.

What the agent builds

A returns and refunds desk with a request flow (order lookup, item selection, reason codes, photo note), and a policy engine where the rules are visible, editable data, auto-approve within 30 days and unopened, flag for review otherwise. The RMA queue moves requests through pending review, approved (with generated RMA number and return instructions), received, refunded, and rejected-with-reason. A metrics strip tracks return rate, top reasons, and this month's refund total; every approved return gets a printable RMA slip.

It seeds twenty orders and requests in every state, so the whole lifecycle is visible on first load.

Make it yours

  • "Match my actual return policy: 14 days, final-sale categories excluded."
  • "Add store-credit as a resolution option alongside refunds."
  • "Track return shipping costs so I can see net loss per return."
  • "Import my orders from a Shopify CSV export."

FAQ

Does it connect to my store? The starting point imports order data; direct integrations are a follow-up conversation depending on your platform.

Why rules-as-data instead of hardcoded? Because policy changes shouldn't need a developer. Edit the rule, the engine behaves differently, and everyone applying the policy applies the same one.

Can customers submit returns themselves? The request flow is exactly that, ask the agent to expose it as a customer-facing page with order-number verification.

Returns & Refunds Desk | OutcomeDev Use Cases