What the documents look like
For after-sales questions in Amazon FBA preparation, this page examines FastGPT multi-turn prompts. Follow planned, in-transit, received and claim records without converting inventory stages into a service conclusion. The examples are hypothetical.
Constraints at the multi-turn prompts step
A proposed inventory exercise lists cartons planned for dispatch and cartons received, which describe different events. Before asking what to do about the missing ones, identify whether the question concerns an undispatched plan, transit awaiting receipt or an observed receiving difference. Preserve period and shipment in the prompt instead of labeling the difference lost. If the answer treats every unreceived carton as compensable, inspect delivery and receiving evidence in input. Without those sources, the claim judgment remains unverifiable.
Configuration and implementation checks
| Conversation check | FBA events and sources | Acceptance clue |
|---|---|---|
| Difference definition | Distinguish plan, transit and receipt | No automatic loss conclusion |
| Added event | Receiving update belongs to the shipment | Actual inclusion in history |
| Question handoff | Identify missing ones and those cartons | Original and optimized question correspondence |
| Application stage | Claim draft versus handling result | Record the unresolved source |
| Business status | Inventory and acceptance records | Conversation does not update records or assign liability |
Documented conditions
| Documented item / implementation check | Condition |
|---|---|
Prompt and task | Define sources, the task and evidence-gap handling. |
Conversation history | Inspect the history actually submitted. |
Question optimization | Review reference resolution and question expansion. |
Model and length | Inspect context and reply length separately. |
Reply format | json_schema needs its format and parsing validation. |
Three common errors
- If the dispatch-plan versus receipt difference becomes lost or compensable stock, inspect shipment, period and carrier events beside the conclusion. Unreceived stock alone establishes no responsibility.
- If a new receiving event still yields an old difference or another shipment, retain update, actual history and optimized question. Distinguish omitted delivery from overwrite after arrival.
- If a claim attachment becomes an accepted or awarded claim, trace the actual handling result. Application wording supplies no such evidence; retain the gap rather than using conversation completion as handling proof.
How to verify the setup
A claim draft may organize known differences and missing documents, but application wording does not establish platform acceptance or responsibility. Record the inventory version and receiving events used at each turn; repeat the question in a new conversation with complete material. An application attachment without a handling result leaves that result unresolved. Mention of compensation cannot promote an application into an award. Confirm the eventual business status against actual inventory and handling records.
| Material context | Scope |
|---|---|
| Document types | FBA Inbound Plan, Carton Label, Inventory Report and Compensation Application |
| Candidate sources | Amazon Seller Central, FBA Inventory Report, Freight Forwarder Tracking and ERP Purchase Order |
| Technical baseline | FastGPT v4.17.2 |