A 25-person startup gets every interviewer to capture the same competencies by doing three things: writing one template per role with four to six competencies and anchored scores, choosing an AI note-taker or interviewer copilot that fills that template during the call instead of producing free-form notes, and requiring independent scores before anyone debriefs. The template is more important than the tool; a great tool with no template produces eloquent notes nobody can compare. This page shows how to design the template, which kind of tool actually fills it, and the consent and retention rules a small company still has to meet.
Designing the template
- Pick four to six competencies the role actually needs. More than six and interviewers skim.
- Write a behavioral anchor for each score. What does a 3 sound like, and what does an 8 sound like? Anchors are what make two interviewers' scores comparable.
- Assign competencies to rounds so each interviewer covers two or three, not all six.
- Add an evidence line per competency and require a quote or a specific example in it.
- End with flags, logistics and a recommendation.
This is the structured interview in practice; the research on why it predicts performance is summarized in structured vs unstructured interviews.
Which tools fill a template
| Tool type | What you get | Fills a template? |
|---|---|---|
| General meeting note-taker | Summary, action items, transcript | No; you retype into the scorecard |
| Meeting note-taker with custom prompts | Summary shaped by a prompt you write | Partly; evidence is often thin and scores are guesses after the fact |
| Interviewer copilot | Questions per competency, live scores with evidence, flags, report | Yes; the report is the template |
| AI interviewer (asynchronous) | Candidate answers to fixed questions, auto-scored | Yes, but no human interviewer in the loop |
On CoPilot Interview, the free Interviewer Mode generates questions per role and scores live; custom question templates per role, shared across the team, are a team-plan feature. The scorecard format itself is in how to write an interview scorecard.
The 25-person startup example
Five people interview: two founders, two leads, one recruiter. They hire for three roles a quarter. Each role gets a template with five competencies; each interviewer owns two. Interviewers run their round with the copilot, which scores the two competencies live, and they add a recommendation line. The recruiter collects the five reports, and the debrief starts with each person's score read out before discussion. Disagreements of more than two points get talked through against the quotes. The whole loop takes one page per interviewer and a 20-minute debrief. Calibration improves within a quarter; how to calibrate interviewers covers the exercise.
Consent and retention at a small company
Small does not mean exempt. Tell candidates in one sentence that an assistant scores answers, keep reports for the same period as other hiring records and then delete them, and keep protected-characteristic information out of the template entirely. If a candidate declines the assistant, score the same template by hand.
Where CoPilot Interview fits
CoPilot Interview's Interviewer Mode is the interviewer-side mode of the app candidates use as an AI interview assistant: role-specific questions, live rubric scoring with evidence notes, flags and a structured report, free on Windows and macOS with no signup. Team plans add custom templates per role and a central evaluation view, which is the setup the startup above needs; see Interviewer Mode. Whatever you use, follow your employer's rules on notice and retention from the first hire.
Disclosure: CoPilot Interview is our product, so weigh what we say about it accordingly.
Frequently asked questions
How do I make every interviewer capture the same competencies?
One template per role with four to six anchored competencies, a tool that fills the template during the call, and independent scores before the debrief.
Can an AI note-taker build a structured interview template?
General meeting note-takers produce summaries, not scored templates. An interviewer copilot scores each competency live with evidence, which is the template filled in.
How many competencies should a template have?
Four to six, each with behavioral anchors for low and high scores, assigned across rounds so each interviewer covers two or three.
Is this overkill for a 25-person company?
No. Small companies feel a bad hire more than large ones, and a one-page template per interviewer with a 20-minute debrief is less work than unstructured notes.
Does CoPilot Interview support templates per role?
Interviewer Mode generates questions per role in the free app; saved custom templates per role shared across a team are part of team plans.
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