What the campaign-setup agent produces, and how a human approves it before anything ships to the platform.
Output should include:
- The full platform-shaped campaign payload (so it can be submitted as-is on approval)
- A human-readable summary of the choices made
- The brief fields that drove each choice (traceability)
- Items flagged as low-confidence or needing decision
- Estimated cost and pacing if calculable
Approval handoff:
- Where does a human see and approve the draft?
- How are edits captured — do they amend the brief or the draft?
- Audit trail entry created on approve / reject / edit
Deliverable: draft-campaign schema plus approval flow.
What the campaign-setup agent produces, and how a human approves it before anything ships to the platform.
Output should include:
Approval handoff:
Deliverable: draft-campaign schema plus approval flow.