Technical reference / quote intake
See where the workflow stops and who moves it forward.
These records come from a LangGraph execution of the five synthetic quote intake cases. Inspect the path, the human review pause and the final handling of each request.
Run evidence
One workflow, five routes
Select a case to inspect its generated execution record.
Recorded graph run / standard-quote
Complete quote request
A known customer requests two products and provides delivery details.
Complete quote request: Approved preview prepared.
Execution path
Steps the graph took
- RequestAccepted
- PolicyPrepared for approval
- ReviewPaused, resumed
- ResultApproved preview prepared
- 01Accept requestpassed
Synthetic request standard-quote accepted.
- 02Check duplicatepassed
No duplicate detected.
- 03Load interpretation fixturepassed
Synthetic extraction fixture loaded; no AI model called.
- 04Apply decision policypassed
approval: Validation passed; a person must approve the prepared output.
- 05Wait for a personpaused
Execution paused at human review; state retained for resume.
- 06Record review decisionapproved
Synthetic reviewer approved the prepared record.
- 07Prepare destination resultpassed
Approved payload preview prepared; no external write was sent.
What this proves
The handoff is part of the workflow.
A TypeScript LangGraph runner uses the same typed fixtures and decision evaluator as the public quote intake demo. It pauses at a reviewer node and resumes with a synthetic decision. The published records are generated from those runs.
This page replays recorded results. It accepts no files, calls no AI model and sends no external message or record. The runner uses an in-memory checkpoint; recovery after a process restart is outside this example.
Check a workflow for your team