← Blog

The Structured Interview Scorecard: A Template That Holds Up

    Most hiring mistakes are not made in the interview. They are made afterward, when three interviewers compare vague impressions and the loudest voice wins. A structured interview scorecard fixes that by making everyone rate the same candidate against the same criteria, on the same scale, with written evidence. It is one of the better-evidenced ways to reduce bias and make hiring decisions you can actually defend later.

    This guide gives you a scorecard template you can copy today, the rules that keep it honest, and the part most teams get wrong: how to capture the interview evidence behind each rating without creating a GDPR or EU AI Act problem for yourself.

    What is a structured interview scorecard?

    A structured interview scorecard is a fixed template that every interviewer fills in for every candidate, rating them against predefined competencies rather than a general gut feel. The competencies are chosen before anyone interviews, the rating scale is the same for everyone, and each rating is backed by a note about what the candidate actually said or did.

    Structured interview scorecard

    A predefined template used by every interviewer to rate a candidate against the same competencies, on the same scale, with written evidence for each rating and a final recommendation. It replaces free-form impressions with consistent, comparable criteria agreed before the interview.

    The value is comparability. When two candidates are scored against identical criteria with evidence attached, the debrief becomes a conversation about the evidence, not a contest of confidence. That is also what makes the decision easier to justify if it is ever challenged.

    The four parts of a scorecard that works

    A usable scorecard is short and specific. Four elements do the work.

    1. Competencies defined up front. Pick the handful of competencies that genuinely predict success in this role, and write them down before interviewing. Resist the generic twenty-item list; four to six real competencies beat a long checklist nobody reads.
    2. A clear rating scale. Use a 1 to 4 scale. An even number of points removes the safe middle option and forces a lean. Define what each number means in one line, so a 3 means the same thing to every interviewer.
    3. A written-evidence field per competency. This is the part that separates a scorecard from a rating. Every score must point to something concrete the candidate said or did. No evidence, no score.
    4. An overall recommendation with a rationale. A single hire or no-hire call with two or three sentences explaining why, tied back to the competency ratings.

    A template you can copy

    Adapt the competencies to your role, keep the structure:

    Competency Rating (1-4) Evidence from the interview
    Role-specific skill   What the candidate said or demonstrated
    Problem-solving   A concrete example from the conversation
    Collaboration   A specific behaviour or story
    Communication   Observed, with a quoted or paraphrased example

    Scale: 1 = clear concern, 2 = below bar, 3 = meets bar, 4 = strong. Finish with one overall recommendation and a short rationale. When you run the debrief, review evidence before revealing ratings, so scores are not anchored by whoever speaks first. Our guide to running a defensible interview debrief covers that sequence in detail.

    The hard part: capturing evidence without breaking GDPR

    The evidence field is where scorecards succeed or fail. If interviewers reconstruct quotes from memory an hour later, the evidence is thin and often wrong. The obvious fix is to record and transcribe the interview so each rating can point to what was actually said. That is the right instinct, and it introduces obligations you have to handle deliberately.

    Interview recordings are sensitive personal data tied to a named individual. To use them as scorecard evidence lawfully, you need a lawful basis, appropriate consent for the recording, a defined retention period, a way for candidates to exercise their rights, and a clear answer to where the audio physically lives. For the consent side specifically, see our guide on recording job interviews in Germany and the EU.

    Residency is the part teams underestimate. Keeping interview audio and transcripts on EU infrastructure removes the cross-border transfer question entirely, which is one fewer thing to defend when a works council or a candidate asks. Where a tool hosts the data is worth confirming in writing before you route real interviews through it; our comparison of EU vs UK data residency for interview notetakers walks through how to check.

    Let the tool capture, keep the judgment human

    There is a strong temptation to let software score the candidate for you. Resist it. Under the EU AI Act, recruitment is a high-risk use case, and a tool that scores, ranks, or filters candidates automatically pulls you into heavier obligations and real fairness risk. A scorecard filled in by humans, supported by a tool that only records and summarizes, sits in a far lighter-touch position.

    That is the line Numi is built on. Numi is an EU-hosted AI meeting assistant used for interviews: it records, transcribes, and summarizes each conversation on European infrastructure and extracts action items and decisions, so every interviewer has an accurate record to ground their rating. It does not score or rank candidates for you, and it does not train on your data. The scorecard, and the decision, stay with your panel, which is exactly where a defensible process keeps them.

    Putting it together

    Write down four to six real competencies, rate each on a 1 to 4 scale, attach evidence to every score, and finish with one recommendation and a rationale. Capture the evidence from an accurate record rather than memory, keep that record on EU infrastructure with consent and a retention schedule, and let humans do the scoring. That is a process that produces better hires and holds up when someone asks how the decision was made. For a shortlist of EU-hosted tools that capture the record without grading people, see our roundup of the best GDPR-compliant AI notetakers for recruiters.

    Frequently asked questions

    What is a structured interview scorecard?

    A structured interview scorecard is a fixed template that every interviewer uses to rate a candidate against the same predefined competencies, on the same rating scale, with written evidence for each rating. It replaces free-form impressions with consistent, comparable criteria agreed before the interview. Structured scoring is one of the better-evidenced ways to reduce bias and improve the predictive quality of hiring decisions.

    What should an interview scorecard include?

    A usable scorecard has four parts: the specific competencies for the role, defined before interviewing; a clear rating scale, typically 1 to 4 with an even number of points to avoid fence-sitting; a written-evidence field for each rating, so a score points to something the candidate actually said or did; and an overall recommendation with a rationale. Keep it to the handful of competencies that genuinely predict success in the role rather than a long generic list.

    Does an AI notetaker create the interview scorecard for me?

    Not with Numi, and that is deliberate. Numi records, transcribes, and summarizes the interview and extracts action items and decisions, so each interviewer has an accurate record of what was said to ground their rating. The rating itself stays with the human. Under the EU AI Act, recruitment is a high-risk use case, and having a tool score or rank candidates automatically pulls you into heavier obligations. Capturing the evidence and leaving the judgment to the panel keeps the scorecard defensible.

    Is it GDPR compliant to store interview scores and recordings?

    It can be, if you handle it properly. You need a lawful basis, a defined retention period, a way for candidates to exercise their rights, and, for recordings, appropriate consent and a clear record of where the audio is stored. Keeping interview audio and transcripts on EU infrastructure simplifies the residency question. Store only the scorecard fields and evidence you actually need, delete on schedule, and confirm where any tool physically hosts the data before you put real interviews through it.

    How do interview scorecards relate to the EU AI Act?

    Using AI in recruitment is treated as a high-risk use case under the EU AI Act, and the classification attaches to what the tool does. A scorecard filled in by humans, supported by an AI tool that only records, transcribes, and summarizes, is far lighter-touch than a system that scores, ranks, or filters candidates automatically. Keeping the decision with the panel and using AI only to capture the record is the cleaner position for most teams.

    Numi is an EU-hosted AI meeting assistant built for interviews. It records, transcribes, and summarizes every conversation on European infrastructure, and never trains on your data.

    Get Early Access