How do you decide whether an agent should ask a human for approval?

Instruction: Explain how you would place human approvals in an agent workflow.

Context: Checks whether the candidate can explain the core concept clearly and connect it to real production decisions. Explain how you would place human approvals in an agent workflow.

Updated

Example Answer

I would require approval when an action has meaningful financial, privacy, or operational consequences, especially when it is hard to reverse. I would decide that through an explicit application policy, rather than the model's own confidence.

For a bulk customer email, the approval screen should show the audience, message, sender, and intended timing. The recorded approval should refer to that exact proposal and have an expiry. If the audience changes, the old approval should not cover the new action.

Before sending, the service still checks the operator's current permission. Approval records consent; it does not create access rights. More approval gates can reduce risk, but excessive gates encourage rubber-stamping, so I would reserve them for meaningful decision points and keep lower-risk work bounded.

Why this works

The example connects risk to an inspectable action and distinguishes consent from the service's authorization decision.

Follow-up to prepare for

How would you prevent approval fatigue? Group related actions only when the complete group is visible and the approval policy permits it.

Your preparation path

Work through these questions in order. Read the answer aloud, then explain it in your own words.

1. Start with the foundations

Build the vocabulary and explain the core decisions.

  1. When would you use a deterministic workflow instead of an agent? Free sample
  2. What state should persist between agent turns? Free sample
  3. How do you decide whether an agent should ask a human for approval? Free sample

2. Apply it to a real workflow

Practice diagnosis, validation, and everyday tradeoffs.

  1. How would you keep an agent from looping on the same failed action? Free sample
  2. When do multi-agent systems create more risk than value? Member answer
  3. An agent keeps choosing the wrong skill. How would you diagnose and fix its catalog? Member answer

3. Prepare for senior discussions

Explain failure boundaries, recovery, and production choices.

  1. Design an agent that can resume after partial failure without repeating side effects. Member answer
  2. After a model upgrade, how would you decide which agent-harness rules to remove? Member answer
  3. An agent's execution container dies halfway through a task. What survives, and how does it recover? Member answer

Related Questions