PRODUCTION AI AGENT // PRE-SCALE CHECKLISTARCHITECTURE✓ Clear agent role boundaries✓ No over-privileged execution role✓ Deterministic stage handoffs✓ Explicit failure states and retriesGOVERNANCE✓ Repo and branch allowlistsQUALITY + EVAL✓ Required test gates✓ Structured evaluator scoring✓ Block / warn / pass thresholds✓ Fallback on failed checksOBSERVABILITY✓ Correlation IDs across all stagesGOVERNANCE cont.✓ Command allowlists for execution✓ Capability flags by environment✓ Global and scoped kill switchesHUMAN REVIEW✓ Explicit merge ownershipOBSERVABILITY cont.✓ Tool-call logging + timestamps✓ Decision logs for evaluator runsROLLOUT✓ Gradual rollout by team / repoScale only multiplies the architecture decisions you already made.

Most agent incidents are not caused by exotic model failure.

They are caused by missing basics.

Use this checklist before scaling any agentic workflow beyond a pilot.

Architecture Checklist

Governance Checklist

Quality and Evaluation Checklist

Observability Checklist

Human Review Checklist

Rollout Checklist

Final Take

If your team can answer “yes” to these items, you likely have a production-ready AI agent foundation.

If not, fix the controls first. Scale only multiplies architecture decisions you already made.