Orchestrating Agent Campaigns with Tickets

Definition

Ticket-based agent orchestration gives every campaign deliverable explicit dependencies, context, checks, ownership, output links and a human-review state so independent work can proceed without hiding blocked decisions.

Perspectives

Shann³ (2026-09-10, X)

Plan the campaign around one shared offer, audience and campaign book, then make vertical briefs for landing pages, paid creative, content, email and outbound. Put each deliverable in a ticket that names prerequisites, files to read, required outputs, checks, status and the exact decision needed when blocked.

Let a coordinating agent assign only ready tickets and let specialist agents revise within a retry or token budget until objective checks pass. Passing checks means ready for human review, not approved to publish. After individual review, check the campaign as a whole for agreement across offer, dates, links and claims; a changed offer sends every affected asset back for revision.

How to apply

• Fits multi-asset campaigns where work can run in parallel but final copy and distribution share dependencies.
• Store the common context in Building a Marketing Agent Workspace and keep each task's scope finite through Operating AI Agents with Bounded Briefs.
• Keep publication behind Guarding Agent Content Publishing with Server-Side Approval.
• Feed campaign outcomes into Closing Campaign Feedback Loops with Data after launch.

Limits

• The source provides a proposed operating design, not measured throughput, error rate or comparison with a single-agent workflow.
• Ticket overhead can exceed its value for a small one-output campaign, and automated checks cannot judge taste or strategic fit alone.

Original article