Reviewing Results

The moment a candidate submits an assessment, Hivemind's AI grades every answer and assembles a full report: scores, written justifications, and behavioral integrity signals. Here's how to read it and how results move candidates through your pipeline.

Open a candidate's report

Get to a report from any of these places:

Assessments → View Candidates: the candidate list for one assessment, with Score and Rank columns (e.g. Top 10%). Click a candidate to open their report.
The candidate profile: assessment results appear alongside everything else you know about them; see Candidate Profiles.
A shareable link: copy the report URL from the report header to share a read-only version with your hiring team.

If grading is still running you'll see Evaluation in Progress; results appear automatically once it finishes, so there's nothing to refresh or trigger.

The Overall Score

The report leads with an Overall Score out of 100 (a weighted average of every question) with a performance badge: Excellent (80+), Good (60+), Average (40+), or Needs Improvement. Below it, a percentile line compares the candidate with others in the same pipeline or with all company candidates who completed the assessment. Click Performance Analytics for a deeper breakdown, or Edit Score to override the overall number for a pipeline candidate.

Questions & Answers

The Questions & Answers section breaks down every response:

Multiple choice: marked Correct Answer or Incorrect Answer, showing the selected and correct options.
Open-ended: the written Response with an AI Score and an AI Grading Justification explaining the mark against the rubric.
Video: the Video Response plays right in the report, with its AI score and justification.
Coding: the Submitted Code in a read-only editor with the language used; live coding also includes the Conversation History with the AI assistant, and a View Session Recording link when available.

Each scored question has its own Edit Score option. Overridden scores are flagged with an Edited pill so your team knows a human adjusted them. Answers that couldn't be auto-graded are called out with the reason. For direct (non-pipeline) candidates, a Retake Assessment button lets you invite a fresh attempt.

Integrity Analysis

Beside the score sits Integrity Analysis: behavioral pattern detection, summarized as a risk Probability, the number of questions Flagged, and a Confidence level. The badge ranges from Standard Patterns through Minor Concerns and Review Recommended up to Requires Attention. Click Detailed Report for per-question analysis of signals like Tab Switches, pasted content, Description Copied, Choice Changes, unusually fast answers, and video Recording Attempts, plus a recommendation for each. A Session Recording replay is available where captured.

⚠️
Integrity signals detect suspicious behavior patterns; they are not proof of cheating, and there is no webcam surveillance involved. Treat Review Recommended and above as a cue to verify skills in a live conversation, not as an automatic rejection.

How results drive stage progression

The overall score is stored on the candidate, and a decision node placed after the assessment stage branches on it: advance high scorers to an interview, reject below your threshold, or route mid-range candidates to manual review. You set those conditions yourself in the pipeline builder; there is no built-in pass mark. Manual score overrides count, so correcting a grade can change where a candidate goes next. See Managing Candidates in a Pipeline.

💡
Cognitive test results show an overall score with Reasoning, Memory, and Verbal components and percentiles; personality reports include role-fit ratings, strengths, challenges, and Evidence from Assessment drawn from the candidate's own answers. See Assessment Types.

What's next

Compare candidates side by side in Managing Candidates in a Pipeline
Move your finalists forward with Offers
Tune weak questions via Creating & Editing Assessments