As of 13 August 2026, AI can only partly score candidates' interview answers.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the role description and the interview plan, then write down the job-relevant criteria and the permitted score for each level.
   2. Remove irrelevant personal information from the transcript where practical, and obtain the candidate's answers in the same format for every candidate.
   3. Paste the role, questions, rubric and one candidate's answers into the prompt, keeping the rubric unchanged between candidates.
   4. Ask the model to produce the evidence, criterion scores, uncertainties and unsupported inferences in the requested table.
   5. Compare every score with the exact answer and rubric, changing any score that lacks quoted evidence or uses a criterion inconsistently.
   6. Have at least one hiring colleague or qualified HR professional compare the candidates' evidence and approve the process before the scores influence a decision.

   Prompt

   Score the interview answers below against the assessment rubric, but do not make the hiring decision.
   
   Role: [ROLE]
   Interview questions:
   [PASTE QUESTIONS]
   Scoring rubric:
   [PASTE THE AGREED CRITERIA AND SCORE DEFINITIONS]
   Candidate answers:
   [PASTE THE TRANSCRIPT OR ANSWERS]
   
   For each criterion, give:
   1. The score allowed by the rubric.
   2. The exact evidence from the candidate's answer.
   3. A short explanation linking the evidence to the criterion.
   4. Any uncertainty or missing evidence.
   
   Do not infer ability, personality, health, disability, age, race, religion, sex, sexual orientation, gender reassignment, pregnancy, marital status or other protected or personal characteristics. Do not score accent, fluency, confidence, style or similarity to the interviewer unless the rubric explicitly makes a job-relevant communication requirement and defines how it is measured. Do not use information outside the answers and rubric. Flag criteria that are ambiguous, unsupported or applied inconsistently. End with a table of criterion scores and a separate list of points that a hiring panel must check before using this assessment. State clearly that the output is a draft assessment for human review, not a hiring decision.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.