A rubric is the difference between two interviewers meaning the same thing by a three and two interviewers arguing about it in the debrief. Most hiring teams know they should have one and few have written the anchors, the sentences that describe what a one, a two, a three and a four actually look like, because writing them is tedious. This page gives copyable anchors for the six competencies most interviews score, communication, problem solving, ownership, teamwork, customer focus and leadership, the evidence that separates a three from a four, how to adapt them per role, and how an interview copilot applies them live. The scorecard that holds them is in the interview scorecard template with AI scoring, and the format they belong to is described in what is a competency-based interview.
Communication and problem solving
| Score | Communication | Problem solving |
|---|---|---|
| 1 | Answers are hard to follow; the point is not clear even with prompting | Describes problems without a method; jumps to a solution or cannot say how it was reached |
| 2 | Clear with prompting; some structure; occasional rambling | Names steps in hindsight; the example is vague about what they did versus what happened |
| 3 | Clear, structured answers that fit the question; adjusts when asked for more or less | A repeatable method shown in one real example: defined the problem, found the cause, chose between options, checked the result |
| 4 | Anticipates what the listener needs; summarises; explains a complex thing simply, unprompted | Method shown in more than one example, with a case where the first approach failed and was changed; names what they would do differently |
The evidence that separates a three from a four in communication is the unprompted summary and the plain explanation of something complex; in problem solving it is the failed first approach, because a candidate who has never had one is either lucky or not describing real work.
Ownership and teamwork
| Score | Ownership | Teamwork |
|---|---|---|
| 1 | Outcomes attributed to others or to circumstances; no example of taking responsibility | Examples are solo; colleagues are absent or obstacles |
| 2 | Takes responsibility when prompted; examples are of assigned tasks completed | Works with others when required; conflict avoided rather than handled |
| 3 | One clear example of owning an outcome beyond the assigned task, including something that went wrong and what they did | Named colleagues in examples; one disagreement handled directly with a result for the work |
| 4 | Pattern of ownership across examples; raised problems early; took on the unowned thing; honest about a failure without blame | Makes others better: credit given, a colleague coached or covered, a disagreement that improved the decision |
The three-to-four line in ownership is the unowned thing: a problem nobody was responsible for that the candidate took. In teamwork it is a story in which someone else got better because of the candidate.
Customer focus and leadership
| Score | Customer focus | Leadership |
|---|---|---|
| 1 | Customers described as a problem; no example of understanding a need | No example of influencing others; leadership described as a title |
| 2 | Service described in general terms; one example without a result | Directed a task when assigned; followers described as resources |
| 3 | One example of understanding a need behind a request and acting on it, with a result and a boundary where one was needed | One example of bringing people to a decision or through a change, with what they listened to and what they held firm on |
| 4 | Pattern across examples; changed something upstream so the problem stopped recurring; honest about a customer they could not satisfy | Developed someone; delivered an unpopular decision well; owned a team failure; evidence that people followed them without a title |
The three-to-four line in customer focus is the upstream fix; in leadership it is someone developed or a failure owned. Adapt the wording per role, a 'customer' for an internal role is the team downstream, but keep the four levels and the behavioural language; calibration between interviewers, as described in how to calibrate interviewers, depends on the anchors being the same text for everyone.
Adapting and applying them
- Pick four to six competencies per role, no more; a rubric with nine is a rubric nobody uses.
- Rewrite the anchors in the role's language, keeping the behaviour: 'resident' for 'customer', 'shift' for 'team', 'patient' for 'client'.
- Add one role-specific competency with its own four anchors, written the same way: safety for a plant, accuracy for billing, judgment for a social worker.
- Give every interviewer the same text, and run one calibration session on a recorded or role-played interview before the first real one.
- Record the evidence sentence for every score; a three without a quote is an impression wearing a number.
- Review the anchors twice a year against hires who worked out and did not; the level that never distinguishes anyone is the one to rewrite.
In Interviewer Mode the rubric appears beside the live transcript, the interviewer scores each answer against the anchors as it is given, and the evidence note is the quote; the report at the end is organised by competency with the scores and the quotes together, which is the scorecard filled in rather than remembered.
Where CoPilot Interview fits
CoPilot Interview is an AI interview assistant for candidates and, in Interviewer Mode, a copilot for interviewers that offers role-specific questions by category, live rubric scoring with evidence notes, follow-up question ideas, red and green flags as hypotheses, and a hire/no-hire report with section scores, flags and quotes, free in the desktop app for Windows and macOS with no signup; the rubric anchors above are the kind of thing its custom templates hold per role, so every interviewer scores against the same text. The candidates on the other side are told, on the candidate side of this site, to follow your employer's rules, and a team that scores against written anchors is holding its own interviewers to a standard it can show.
Disclosure: CoPilot Interview is our product, so weigh what we say about it accordingly.
Frequently asked questions
What is an interview rubric?
A set of competencies with written anchors describing what each score level looks like in behaviour, so different interviewers mean the same thing by the same number.
How many competencies should a rubric have?
Four to six per role, each with four anchored levels. More than that and interviewers stop using it.
What separates a three from a four?
Evidence of a pattern rather than one example, and the harder behaviours: the failed first approach, the unowned problem, someone made better, the upstream fix, a failure owned.
How do I adapt these anchors to my roles?
Keep the levels and the behavioural language, swap the nouns for the role's, add one role-specific competency written the same way, and calibrate once before using it.
How does an interview copilot use a rubric?
It shows the anchors beside the live transcript, the interviewer scores each answer against them with an evidence quote, and the report is organised by competency with scores and quotes together.
Try Interviewer Mode on your next screen
Role-specific questions, live rubric scoring with evidence notes and a structured report at the end of the call. Free in the desktop app for Windows and macOS, no signup.
Get the desktop app