Home · Business · HR & People · Interviewing & assessment
As of 13 August 2026, AI can only partly make a hiring decision from interview results.
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 purpose-built recruitment software for candidate scoring, enrichment and pipeline automation, while a human-led hiring process has no fixed cost stated here.
If this goes wrong: you reject a suitable candidate or select an unsuitable one, and the decision may create an unfairness complaint or employment dispute.
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 job description and write a short list of the essential and desirable criteria, separating job-related requirements from preferences.
- Ask every interviewer for their notes and scores, then remove irrelevant personal details and label each item by candidate, question and interviewer.
- Gather the agreed interview questions, scoring guide, work-sample results and any notes explaining reasonable adjustments or inconsistent questioning.
- Paste the criteria, scoring guide and anonymised evidence into a chatbot using the prompt above, asking it to mark missing evidence instead of filling gaps.
- Compare the chatbot's scores and provisional ranking with the original scoring guide and the source interview notes, correcting any claim that is not supported.
- Ask a colleague involved in the process to check that candidates were assessed against the same criteria and that no protected characteristic or unrelated personal factor influenced the comparison.
- Record the human panel's final decision, reasons, dissenting views and evidence considered, then send the outcome through your normal recruitment process.
Prompt
You are helping with a UK hiring process. Do not make an automatic or final hiring decision, and do not use protected characteristics or proxies for them. Compare the candidates only against the job-related criteria below. Job title: [JOB TITLE] Job criteria, with importance for each criterion: [PASTE THE AGREED CRITERIA] Interview questions and scoring guide: [PASTE THE QUESTIONS AND SCORING GUIDE] Candidate interview evidence, labelled by candidate and question: [PASTE STRUCTURED NOTES] Any work sample or other evidence: [PASTE EVIDENCE OR WRITE NONE] For each candidate: 1. Summarise only evidence actually present in the material. 2. Score each criterion using the supplied scoring guide, marking insufficient evidence as "not evidenced" rather than guessing. 3. Separate evidence from interpretation and identify contradictions between interviewers. 4. Identify questions that were not asked consistently or criteria that were applied inconsistently. 5. Flag any wording that could indicate bias, reliance on a protected characteristic, or a factor unrelated to the job. 6. Explain what further evidence would be needed before deciding. 7. Give a provisional ranking only if the evidence supports it, with reasons tied to the agreed criteria. End with a short decision record for a human hiring panel. State that the panel must make and record the final decision, check the process for consistency and discrimination, and keep the underlying evidence. Do not invent facts, scores or legal conclusions. This is not professional advice; refer a serious employment-law issue to an employment solicitor or qualified HR professional.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- AI cannot know whether an interview criterion is genuinely necessary for the role or an unconscious preference disguised as a requirement.
- AI cannot reliably detect every discriminatory assumption in notes, scoring systems or interview conduct.
- AI cannot resolve conflicting evidence by speaking to interviewers, checking references or assessing a candidate in person.
- The hiring manager remains accountable for the decision, its consistency and any complaint or employment dispute.
- Pasting interview notes into an external tool can create confidentiality and data-protection risks unless your organisation permits it and the data is handled appropriately.
What caps this at PARTLY: legal accountability, judgement under ambiguity and stakes of error.
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 | 1 |
| Verification | 1 |
| Liability | 0 |
| Effort delta | 1 |
| Total | 5 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI make a hiring decision from interview results?
- AI can compare structured interview evidence with agreed job criteria and give you a provisional ranking. It should not make the final decision because the employer remains accountable for a fair, consistent and lawful process.
- Is it legal to use AI to decide who gets a job in the UK?
- Using AI in recruitment does not remove your duties under UK equality and employment law, and an automated recommendation can still produce discriminatory outcomes. This is not professional advice; refer a serious employment-law issue to an employment solicitor or qualified HR professional.
- Can AI score interview answers?
- Yes, if you provide a clear scoring guide and structured notes, AI can organise evidence and apply the stated criteria. It cannot reliably judge missing context, interviewer bias or whether the criteria themselves are appropriate.
- Should I upload interview notes to ChatGPT?
- Only if your organisation allows it and you have a lawful, secure process for handling the information. Remove unnecessary personal data, check the tool's approved-use rules and treat the output as an aid for a human panel, not as the hiring decision.
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.