Integrity

How to Read an Integrity Report After a Remote Technical Interview

ClarityHire Team(Editorial)4 min read

What an integrity report actually is

An integrity report is the post-interview artefact that records what the platform observed during a remote assessment or live interview — paste events, tab switches, face-presence anomalies, audio-video desync, code-coherence anomalies, and so on.

It is not a verdict. The single biggest mistake hiring managers make is treating the report's top-line "severity: critical" badge as a reason to reject. That's not what the signal is for. The report's job is to surface anomalies; your job is to interpret them.

Read in this order

1. Start with the timeline, not the summary. A summary that says "12 anomalies, severity: warning" tells you nothing. Open the timeline view and look at when the anomalies happened. Anomalies clustered at minute 4 of a 30-minute round are very different from anomalies spread evenly across the session.

2. Cross-check with the work product. A "tab switch" event during a coding question is suspicious. A tab switch at minute 28, after the candidate has already submitted, is irrelevant. Always pair each integrity signal with what the candidate was doing at that timestamp.

3. Categorise the signal. There are three classes that matter:

  • Strong signal — large paste of the exact answer, multi-person face detection, a long audio dropout that coincides with the candidate "finishing" a problem.
  • Soft signal — a single tab switch, a brief gaze drop, a 200-character paste that could be boilerplate.
  • Environmental noise — webcam reflection mistaken for a second face, autocomplete-driven paste, a roommate walking past.

Treat strong signals as worth a follow-up. Treat soft signals as worth noting in the scorecard, not as standalone evidence. Treat noise as noise.

What the report can never tell you

The integrity layer deliberately does not store biometric templates — no face recognition, no keystroke biometric profiling against a stored print. That means the report can flag "two faces visible" but cannot say "this is the same person as last time" or "this is the candidate from the CV". Identity verification is a separate process and shouldn't be inferred from integrity data.

It also can't tell you intent. A long paste might be cheating; it might also be the candidate copying a snippet they wrote in their preferred editor. Always ask before assuming.

The follow-up pattern

When you see a strong signal, don't reject — ask. A 5-minute follow-up with the candidate solves more cases than a unilateral decision ever will:

  1. "At minute 14 the platform flagged a large paste. Can you walk me through where that came from?"
  2. "You had a 90-second audio dropout starting at minute 22 — what happened on your end?"

A real candidate explains and produces evidence. A cheating candidate stalls, deflects, or produces a story that doesn't match the timeline. The conversation does more work than any score.

What to record in the scorecard

Treat the integrity report as one input alongside the structured rubric and the interviewer's written feedback. In the candidate's record, write:

  • Which specific signals fired (not the summary badge).
  • Whether you ran a follow-up, and what the candidate said.
  • Your final read: confirmed concern, addressed, or noise.

This trail protects you if the candidate appeals — and forces honest reasoning at the moment of decision, which is when bias is most likely to creep in.

How ClarityHire structures the report

Every assessment and interview produces a timeline view with paste events, tab/focus changes, face-presence anomalies, code-coherence flags, and keystroke-rate outliers. Severity badges are advisory — every signal is clickable down to the exact second, and every flag includes the raw underlying event so reviewers can judge it themselves. No auto-rejections fire from integrity data alone.

TL;DR

An integrity report is evidence, not a verdict. Read the timeline before the summary. Distinguish strong signals from soft signals from noise. Always follow up with the candidate before deciding. The point of the report is to give you a basis for an honest conversation, not to outsource the decision to the platform.

integrity reportremote interviewcheat detection signalsinterview reviewfair hiring

Related Articles