Draft an interview record with the interviewer’s assessment

Why use this

1. Interview notes need to preserve what was said while recording the interviewer’s own evaluation in the required form. A transcript alone cannot supply that judgment, and a polished summary can accidentally make tentative impressions sound like established conclusions.

2. This job could reduce transcription and form-mapping work while leaving assessment ownership visible. A review draft that distinguishes candidate statements, transcript evidence and interviewer comments gives the hiring team a clearer basis for correcting the record before submission.

Why an agent / simpler alternative

Useful judgment

Map varied conversation topics into role-specific fields and ask targeted questions where human assessment is missing.

Simpler alternative

Use structured interviewer notes when the form is short and completed during the conversation.

Unsuitable when

The agent is expected to rank candidates or infer protected or sensitive traits.

Proposed inputs

1. Name

   Interview transcript

   Owner

   Authorized hiring team

   Minimum scope

   Consented transcript with speaker labels and interview date

   Quality check

   Check transcription uncertainty and exclude unrelated personal material.

2. Name

   Role form and assessment

   Owner

   Interviewer

   Minimum scope

   Current form fields plus the interviewer’s explicit judgments

   Quality check

   Do not substitute transcript sentiment for assessment.

Access and setup — proposed or unknown

Proposed minimum

1. Read a scoped transcript and blank form; prepare a private submission preview.

Compatibility unknowns

1. Unknown: Granola export quality, recording permissions and ATS field compatibility.

External actions requiring approval

1. Submitting or revising the ATS record and sharing candidate information.

Expected deliverable and review checks

Proposed artifact

Draft role form with evidence excerpts, interviewer assessment, unanswered fields and a submission review status.

Critical checks

1. Ratings come from the interviewer, not invented inference.

2. Each assessment is attached to the correct candidate and role.

Missing output evidence

This overlay adds no new output inspection or execution validation. Preserve the canonical implementation evidence and its outcome limits.

Failure signals and proposed responses

1. Signal

   A transcription error changes a technical answer.

   Why it matters

   The record may unfairly misrepresent the candidate.

   Proposed response

   Mark the passage uncertain and ask the interviewer to verify it against authorized material.

2. Signal

   The form requires a rating that the interviewer never supplied.

   Why it matters

   Autofilling creates a judgment no person made.

   Proposed response

   Leave the rating pending and request the missing assessment explicitly.

Effort and cost — unknown

Setup work

1. Map transcript topics to the current role form and define attribution labels.

Ongoing review

1. Review uncertain quotes and obtain missing judgments.

Unknowns

1. Unknown: transcription costs, ATS access and interviewer review time.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

A synthetic interview includes a garbled answer, a clear example and an unanswered rating field.

Acceptance checks

1. The clear example is mapped with attribution.

2. The garbled answer and rating remain unresolved until human input.

Stop conditions

1. Candidate scoring, protected-trait inference or real ATS submission is required.

Limits

1. Interviewers own assessments. Submission review is a proposed safeguard, not a claim about the original run.