Design retry and idempotency controls for AI-triggered actions.

Instruction: Describe how you would make AI-triggered actions safe to retry.

Context: Assesses whether the candidate can design a practical architecture and explain the main tradeoffs. Describe how you would make AI-triggered actions safe to retry.

Official answer available

Preview the opening of the answer, then unlock the full walkthrough.

I would design retries as a platform capability. The assistant can request an action again, but the system should decide...

Upgrade to view official answer

Related Questions