E-commerce After-Sales: Turn Return/Exchange Messages into Processing Tickets
Extract order, issue, and missing materials, then form a pending-review after-sales ticket and reply draft according to store rules.
Fictional store: 森叶家居 Order: EC-2041, delivery date 2026-09-12, inquiry date 2026-09-15 Customer message: The white cup received has a crack at the rim; I want to exchange it. Photos will be added later. Demo rule: For damage issues, first collect product photos and packaging photos, then customer service decides the handling plan.
Read both input and output
Input
Fictional store: 森叶家居 Order: EC-2041, delivery date 2026-09-12, inquiry date 2026-09-15 Customer message: The white cup received has a crack at the rim; I want to exchange it. Photos will be added later. Demo rule: For damage issues, first collect product photos and packaging photos, then customer service decides the handling plan.
Output
Preset ticket Order: EC-2041 Issue category: damage on arrival; Customer intent: exchange Missing materials: product damage photo, packaging photo Next step: customer service collects additional evidence; do not yet promise a successful exchange Reply draft: Hello, the rim damage has been recorded. Please provide photos of the product and outer packaging; we will explain the handling plan after verification. Execution status: No refund issued, no exchange order created, no reply sent.
Applicable Scenarios
Extract order, issue, and missing materials, then form a pending-review after-sales ticket and reply draft according to store rules.
This demo uses fictional business data and pre-prepared outputs to show a customizable workflow; it is not a customer case, live AI execution, or actual business result.
Work Steps
- Identify the order number, item, issue, and customer intent from the message.
- Check whether materials are complete against the provided after-sales rules; escalate to a human when rules do not cover the case.
- Generate ticket fields and a reply draft for customer service to confirm.
Human Review and Boundaries
Customer service verifies the actual order, after-sales policy, and image evidence before deciding whether to refund or exchange. Customer statements must not be directly treated as a conclusion on product quality.
Customizable Elements
Customize after-sales categories, store scripts, and ticket fields; according to agreement, connect to order and customer service systems, with review checkpoints for write operations.
How to Validate
Test separately with normal inputs, missing fields, and conflicting information; check that key fields can be traced to the input, missing content is clearly marked, and numerical calculations are completed by reviewable rules. The samples and delivery scope agreed by both parties shall prevail.
Platform and Delivery
Can customize instructions, templates, supporting scripts, and review steps for Codex or Claude. Specific product versions, tool permissions, and system interfaces are confirmed during the requirements phase and tested separately. Deliverables can include source files, test samples, and usage documentation; this page does not connect to external systems, send notifications, or perform business operations.