What you do
- Provide the offer, audience hypotheses, destination, brand constraints, and channel preference.
- Approve generation cost when new assets are needed.
- Review the candidate set and approve any provider upload separately.
Growth and creative teams developing paid-media variants / media buying
A focused candidate set tied to an experiment hypothesis, with generation and provider upload kept behind separate approvals.
The problem
AI can explore several evidence-based variants quickly while reviewers preserve brand, claim, and experimental discipline.
What you do
What the agent does
When this workflow helps
Before it starts
What proves it worked
Workflow path
A real run expands these stages around the request, context, selected tools, approvals, and dependencies.
Read recent creatives, learnings, active experiments, and constraints.
Create explicit candidate angles, copy needs, asset needs, and review criteria.
Use granted utility actions or existing asset refs to create the candidate asset set.
Persist reviewed candidate creative records and their provenance.
Execute approved provider upload or campaign update actions only when granted by the concrete run plan.
Safe stopping and recovery
Documented follow-on work
The candidates require a measured creative-and-page experiment.
conditionalMedia Buying Campaign LaunchThe candidates belong to a new approved campaign launch.
Specialists inside this workflow
Connected work