The Deployment & Rollout Plan
1 fill-in slot · from The AI Prompt Handbook for AI Agents & MCP
Your details
0/1 filledThe Deployment & Rollout Plan
Plan the rollout of this agent to production. System: [DESCRIBE] Current state: [DESCRIBE] Users: [DESCRIBE] Risk: [DESCRIBE] Produce: 1. Rollout stages, from safest to full 2. What is validated at each stage before proceeding 3. Success criteria per stage — specific and measurable 4. Rollback criteria per stage — the signal that stops the rollout 5. What runs in shadow mode first, comparing against current behaviour without acting 6. Feature flag design: what can be turned off independently 7. Monitoring specific to the rollout 8. Communication: what users are told, and when 9. What is the smallest possible first deployment that would teach me something real? Then: 10. What could go wrong during rollout that would not show in testing? 11. What is the fastest possible path from "something is wrong" to "it is off"? For a system with side effects, shadow mode is usually worth the effort. Tell me specifically what shadow mode would look like here.
Highlighted [slots] are still empty. Downloads are plain .txt — paste into any AI assistant.