Reading the interview results page
When a candidate completes their AI interview, you get an Interview Review page with the AI evaluation, transcript, recordings, and decisions. The full layout, explained panel by panel.
Key takeaways
- The page has four major sections: AI Evaluation, Response Metrics, Strengths/Concerns, and the decision bar.
- The header has shortcuts to view the full report, download the PDF, open the candidate, and open the job.
- Decision buttons at the bottom: *Reject*, *Schedule next round*, *Shortlist*.
- Your decision is recorded on the candidate, not on the interview — meaning future interviews for the same candidate see your prior decision.
Before you start
When a candidate finishes their AI interview, you get an Interview Review page with the AI's full evaluation, the transcript, recordings, and controls to advance or reject. This article walks through every section so you know exactly what each one is telling you.
How to get here
Two paths:
- Sidebar → Interviews, find the row, click View.
- Dashboard → Interviews ready to review (top KPI) → Review now → click any row.
Either way you land on this page. The candidate name shows in the breadcrumb and again as the page subtitle.
Header — quick actions
Four buttons at the top right:
- View Report — opens a printable/sharable summary view of just the evaluation.
- Download PDF — exports the evaluation + transcript as a PDF for offline sharing.
- View Candidate — jumps to the candidate's full profile (their resume, other jobs they're linked to, all their interviews).
- View Job — jumps to the job this interview was conducted for.
Next to the candidate's name is the interview status — for a completed interview, you'll see Completed in a green badge.
Section 1 — AI Evaluation
The big panel on the left. The headline is the overall recommendation: Strongly Recommend, Recommend, Hold, Not Recommended, or Strongly Not Recommended. Treat this as the AI's summary judgement, akin to a recommending interviewer's bottom line.
Below the recommendation, the six dimension scores:
- Technical Competence — depth on role-relevant skills (weight varies by role).
- Communication — clarity, structure, and depth of explanation.
- Experience Relevance — alignment between the candidate's background and the JD.
- Problem Solving — reasoning quality on open-ended questions.
- Cultural Fit — how the candidate describes work, conflict, growth.
- Interview Performance — pace and structure overall.
Each score is 0–100, and the percentage next to it (e.g. 30%) shows the weight that dimension carries in the overall recommendation for this role. Weights are role-aware — Technical Competence is weighted more for engineering roles than for sales.
TIP
If a candidate has a strong overall score but a weak dimension you care about, drill into that dimension in [Reading the evaluation report](/knowledge-hub/interviews/evaluation-report) before making a decision. A great engineer with poor communication may still be hireable; it depends on the role's seniority and team norms.
Section 2 — Response Metrics
A row of objective signals from the audio analysis:
- STAR Compliance — how often the candidate's answers followed the Situation/Task/Action/Result structure. High is good (they organise their thinking) but extreme highs can mean rehearsed answers.
- Avg Response — average length of an answer, in words. Very short answers usually mean the candidate did not engage; very long ones (1000+ words) can mean rambling.
- Filler Words — percentage of um, uh, like, you know. Useful as a confidence signal but contextual — some candidates with strong substance have high filler counts, and that is fine.
- Confidence Trend — Stable, Rising, Declining. Whether the candidate sounded more or less confident as the interview progressed. Declining can indicate a candidate who struggles when questions get harder.
Section 3 — Strengths & Areas of Concern
The AI writes a short list of specific things the candidate did well and specific things that gave it pause. These are the most useful inputs for a hiring debrief — much more actionable than the numeric scores alone.
Each bullet references the actual conversation. If the AI says "Provided quantified achievements (73% cost reduction in logistics ops)", you can search the transcript for that figure and verify.
Section 4 — Interview Details (right panel)
Metadata about the interview itself:
- Candidate — name and contact email.
- Position and organisation the interview was conducted for.
- Invited — when you sent the invitation.
- Duration — how long the actual interview ran (may be shorter than the configured slot if the AI finished early).
- Completed — when the candidate clicked done.
Section 5 — Decision bar (bottom)
The buttons that move the candidate forward (or not):
- Reject — closes the candidate for this role. They are not auto-emailed; you can send a rejection separately.
- Schedule next round — advances them to the next hiring stage. If you have follow-up rounds configured, this can trigger them.
- Shortlist — marks the candidate as a favourite without advancing. Useful for parking a decision overnight.
Your decision is recorded on the candidate within the context of this job. The same candidate can be on different paths in different jobs — Rejected for Engineer X, Shortlisted for Engineer Y.
WARNING
*Reject* is reversible — you can re-open the candidate from their profile — but it is a strong signal to your team. Use *Hold* (or just don't decide yet) if you're unsure, rather than rejecting and reversing later.
Next: a deeper look at how to read the evaluation report itself.
Frequently asked questions
Can I see what questions the AI asked?
Yes. Open the full transcript (button in the header). Questions are clearly delineated from answers, with timestamps. The questions themselves come from the interview *plan* generated for this role from the JD and any custom configuration.
The AI gave a strong recommendation but I disagree. What should I do?
Trust your judgement. The AI evaluation is one signal — you should give it serious weight, but you can absolutely override. Click *Reject* or *Hold* even if the AI said *Strongly Recommend*. The platform records both the AI recommendation and your override so you can review your team's alignment over time in [Analytics](/knowledge-hub/analytics).
The scores look too low or too high — is the AI calibrated correctly?
AI scores are calibrated against a benchmark of similar role-archetype interviews. Within a single role, scores are comparable to each other (an 80 in one Operations interview is genuinely better than a 60 in another). Across very different roles (Engineer vs Sales), direct numeric comparison is less meaningful — use the *Recommendation* field instead.
What's the difference between *Schedule next round* and *Shortlist*?
*Schedule next round* moves the candidate to the next stage in your hiring funnel and (if you have it configured) triggers the next-round invitation. *Shortlist* just flags the candidate as a favourite without advancing them — useful when you want to think on a decision overnight.
Next steps
Reading the AI evaluation report — what the scores actually mean
Six dimensions, weighted by role, summed into one recommendation. Here is how to read each score, when to trust it, and when to override it with your own judgement.
Reading the transcript
The transcript is the raw record of the conversation. Here is how it's structured, what the timestamps mean, and how to use it alongside the AI evaluation.