One interview scorecard for the whole hiring team, with AI

What the scorecard holds

1. Competencies: four to six, each from an outcome or must-have in the role, each owned by one interviewer.
2. Anchors: 1 Poor, 2 Borderline, 3 Solid, 4 Outstanding, each written as what the answer contains. Level 3 is the bar for the role.
3. Evidence fields: the question asked, what the candidate said, the rating. A rating with no evidence doesn't count.
4. Rating rules: alone, right after the interview, before any discussion. "Not assessed" is allowed and isn't a 1.
5. The hire rule, written and frozen before the first interview.
6. Job-related only: no field for age, family, health, nationality, religion or appearance.

A worked example

Illustration: a made-up business. A 30-person company is hiring an operations coordinator. The hiring manager gives the AI the role scorecard and gets a draft with five competencies, one per interviewer. One of them:

The hire rule, frozen before the first interview: hire only if every must-have is 3 or higher across the interviews, with no 1 anywhere and at least one 4; any 2 on a must-have is a no-hire unless the debrief names evidence that resolves it. The manager deletes one competency the draft proposed, "culture fit", because it had no definition anyone could observe.

Roll it out to the hiring team

1. Build the scorecard from the role, not from last year's interview questions.
2. Give each interviewer their competencies and questions before the interviews.
3. Collect every rating before the debrief starts; discuss the biggest disagreement first.
4. Keep the rule frozen: change it for the next role, not mid-loop.
5. Share the scorecard skill so every hiring manager's AI drafts the same way.

The EEOC's rules on what may be asked apply in the US; in Japan, see the Japanese version of this page.

Next step: the rest of the hiring skills

The scorecard is one of five skills a hiring team uses. Claude skills for recruiting covers the others, from the role intake to reference checks. For the questions themselves, see interview questions with ChatGPT.

Questions

What is an interview scorecard?

One sheet every interviewer rates a candidate on: four to six competencies from the role, a 1 to 4 scale with each level described as what an answer contains, a field for what the candidate said, and the rule for a hire. Built before the first interview, it lets the team compare candidates on the same evidence.

Should interviewers discuss a candidate before scoring?

No. Canada's Public Service Commission recommends that board members rate independently shortly after each interview, then reach a group consensus. Rating alone first keeps one confident voice from setting everyone's score.

Can AI write the scorecard?

It can draft one from the role scorecard: competencies that can be observed in an answer, anchors for each level, evidence fields and a hire rule. The hiring manager checks every competency against the role and freezes the rule before the first interview.