# AI interviews: working checklist

Companion to https://lunaprompts.com/blog/ai-interviews-campus-hiring-guide

LunaPrompts Editorial. Original illustrative exercise.

## Define the workflow

- Role or campaign:
- Owner:
- Decision this workflow should support:
- Candidate instructions and support route:

## Worked example: Design a useful graduate screening round

For a graduate engineering campaign, define an early-round objective: understand what the candidate personally contributed to a project and whether they can explain a technical choice. Those are narrower and more reviewable aims than “assess overall potential.”

Use a common framework with room for relevant follow-ups. Start with the project’s purpose, ask about one piece the candidate owned, and explore a failure or difficult decision. Review answers against the same role-related criteria, while allowing the examples to differ.

Tell candidates that the conversation is AI-led, what it covers, and how to reach the recruiting team if the session fails or they need an alternative arrangement. Include this information in the invitation rather than making it a surprise inside the interview.

Pilot the workflow before a large campaign. Check whether the questions are understandable, whether the output preserves useful evidence, and whether reviewers can distinguish a demonstrated skill from an unanswered question. Reserve human review capacity for the results. The next interviewer should receive a concise handoff with evidence and a proposed topic, not only a generic recommendation.

## Review checklist

- [ ] Purpose: The question this round must answer
  Evidence or open question:
- [ ] Framework: Common topics and relevant follow-ups
  Evidence or open question:
- [ ] Candidate experience: AI disclosure, instructions, and help route
  Evidence or open question:
- [ ] Handoff: Evidence, uncertainty, and human review owner
  Evidence or open question:

## Working session

Write the first-round objective in one sentence. Remove any interview topic that does not help answer it.

## Avoid this mistake

Interview volume is not a quality measure by itself. Review whether the conversation supplies useful evidence and whether candidates can complete it as intended.

## Record the next step

- Observed evidence:
- What remains uncertain:
- Person responsible:
- Next action and date:

Keep candidate examples anonymous when sharing this worksheet.