Home · Business · HR & People · Job ads & screening
As of 13 August 2026, AI can only partly create an interview scorecard.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ita colleague
What the alternative costsManatal is described as recruitment software with candidate scoring, enrichment and pipeline automation, making it a direct software alternative to a hand-built scorecard.
If this goes wrong: interviewers score candidates against vague, inconsistent or indirectly discriminatory criteria and the employer faces a complaint or makes a poor hire.
What to actually do
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.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Open the current job description and copy the role title, level, duties, essential criteria and desirable criteria into the prompt.
- Gather the interview stage, interview length, interviewer roles and any competencies already approved by your organisation, then paste them into the matching slots.
- Paste the completed prompt into ChatGPT, Claude or Gemini and ask it to generate the scorecard using the specified 1 to 5 scale.
- Compare every competency and question with the approved job description, removing anything that does not relate directly to the work.
- Ask an HR colleague to check the fairness check, reasonable-adjustment wording and any flagged employment-law issue before the scorecard is used.
- Run the same questions and scoring rules past every interviewer, then save the approved version in the organisation's recruitment records.
Prompt
Create an interview scorecard for the UK role described below. Role and level: [insert role title and level] Job description: [paste the full job description] Essential criteria: [list the criteria] Desirable criteria: [list the criteria] Interview stage: [insert stage and duration] Interviewers: [insert interviewer roles] Competencies to assess: [list them, or derive them only from the job description] Scoring scale: use 1 to 5, with clear behavioural anchors for each score Produce: 1. A table of competencies, weighting, what good evidence looks like, and what should not be scored. 2. Two structured interview questions for each competency, with one follow-up probe for each question. 3. A 1 to 5 scoring guide based on evidence in the candidate's answer, not confidence, accent, personality, background or similarity to the interviewer. 4. A final score calculation that separates essential criteria from desirable criteria and explains how an interviewer should record evidence. 5. A short instruction sheet for interviewers explaining how to ask the same core questions, avoid leading questions, record evidence and handle reasonable adjustments. 6. A fairness check identifying any criterion that is vague, unrelated to the role, likely to disadvantage a protected group, or difficult to assess consistently. Use only information supplied above. Do not invent duties, qualifications or legal conclusions. Do not recommend scoring protected characteristics or personal circumstances. Flag any point that needs review by a UK HR professional or employment solicitor. Keep the language plain and make the scorecard practical to use in one interview.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- AI cannot decide which criteria are genuinely necessary for the job when managers disagree or the role is poorly defined.
- AI cannot guarantee that a scoring system will not disadvantage a protected group or create an indirect discrimination risk.
- AI cannot replace the employer's decision about how much evidence is enough to appoint someone.
- AI cannot make interviewers apply the scorecard consistently or stop them relying on unrecorded impressions.
- AI cannot take responsibility for an unfair process, rejected candidate or unsuitable hire.
What caps this at PARTLY: legal accountability, judgement under ambiguity and verification cost.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 0 |
| Effort delta | 2 |
| Total | 7 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can ChatGPT create an interview scorecard?
- Yes, it can draft a useful scorecard from a job description, including competencies, structured questions, evidence prompts and rating anchors. Treat the draft as a starting point and have an HR colleague check fairness and UK employment-law risks before use.
- What should be included in an interview scorecard?
- Include job-related competencies, consistent questions, evidence-based scoring anchors, weightings and space for notes. Do not score accent, confidence, personality, personal circumstances or anything unrelated to the role.
- Is an AI-generated interview scorecard legal in the UK?
- An AI-generated scorecard is not automatically lawful or unlawful, but the employer remains responsible for how it affects candidates and decisions. This is not professional advice; a serious equality or recruitment-process concern needs review by a UK HR professional or employment solicitor.
- How do I check an AI interview scorecard?
- Compare every criterion with the approved job description and ask an HR colleague to check for vague requirements, inconsistent scoring and possible disadvantage to protected groups. Test that different interviewers would reach similar scores from the same evidence before using it with candidates.
Nearby answers
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.