LunaPromptsTwo candidates score 82. The recruiter needs to prepare the next interview, but the number does not say whether to explore a missed edge case, an unclear design decision, or a debugging gap.
A useful report earns its space by making the next question more precise. The score helps you navigate. The submitted work and supporting evidence explain what is worth discussing.
LunaPrompts combines assessment results with candidate telemetry, available recordings, and AI-assisted evaluation. Recruiters and hiring managers can review demonstrated skills, inspect how work developed, and decide what needs further discussion.
Compare two similar assessment scores
Hiring managerIllustrative scenario01 / 03Both candidates score 82
This fictional comparison starts with the same total and different evidence.
In this fictional comparison, both candidates have a total of 82. Candidate A’s implementation satisfies the core requirement but fails on an empty collection. Candidate B’s implementation passes the supplied examples, while their explanation does not identify why an earlier version failed. These are different uncertainties.
For A, ask what the interface promises for empty input and request a small test. For B, show the earlier failure and ask for a diagnosis linked to a specific line or state change. Neither follow-up needs to become a second full assessment.
Write report notes in an evidence-first order: observation, relevance to the role, unresolved question, proposed follow-up. “The function raises an error for empty input” is an observation. “Careless engineer” is a broad personal judgment the example does not support.
Use telemetry to locate the relevant moment, not to invent the candidate’s internal reasoning. Someone may pause to read documentation, use an accessibility tool, or consider a design. The assessment should invite an explanation when that context matters. If the available recording or event history does not resolve a question, the next interview can.
Start with the role, then open the evidence
Skill-level results help you distinguish a candidate who is strong in coding but needs support with system design from one who shows the opposite pattern. Cohort comparisons add context, while the role's requirements determine which strengths matter most.
Use the overview to find the relevant questions. Open the candidate report to understand the submission, the section results, and the available supporting evidence.
What telemetry adds to a candidate report
Depending on the assessment environment and enabled collection, telemetry can include attempts, edits, typing timing, paste patterns, test runs, and session events. Recordings can provide additional review context where that feature is enabled.
These details help reviewers reconstruct observable work. They do not reveal a candidate's private thoughts, and typing speed is not a direct measure of engineering ability.
| Evidence | A useful review question | What it cannot establish alone |
|---|---|---|
| Passed and failed tests | Which requirements and edge cases did the solution satisfy? | Overall suitability for the role |
| Attempts and revisions | How did the implementation change after feedback? | The candidate's intent or every step of their reasoning |
| Typing and paste events | Which parts of the timeline need more context? | Misconduct or skill based on speed alone |
| Code and style review | Is the solution understandable and maintainable for this task? | How the candidate would perform in every codebase |
| Recordings and session events | What happened around a flagged or important moment? | A conclusion without considering context |
| Candidate explanation | Can the candidate justify and adapt the solution? | A substitute for checking the submitted work |
Use AI Judge as a review copilot
AI Judge helps interpret candidate evidence against the role and its evaluation criteria. Its contribution is a review aid: a way to surface strengths, gaps, and questions the hiring team can examine.
The reviewer should be able to connect an assessment judgment to relevant work. Where the evidence is incomplete or inconsistent, that uncertainty is useful information for the next interview. The hiring team remains responsible for deciding who progresses.
Help the whole hiring team share context
Recruiters need a clear explanation of why someone should move forward. Interviewers need to know what has already been tested. Hiring managers need to understand the evidence and its limits.
A useful report connects these needs. It should keep results, work samples, and unresolved questions close enough that the team can discuss the candidate without reconstructing the process from scattered notes.
A review framework your team can use
The following is an editorial checklist for this workflow. Adapt it to the role, the candidate journey, and your configured environment.
| Review area | What to establish |
|---|---|
| Observation | A concrete result, artifact, or event |
| Role relevance | Why this requirement matters |
| Uncertainty | What the evidence does not settle |
| Follow-up | The smallest useful next question |
More telemetry does not automatically mean more understanding. Avoid turning typing speed, pauses, or paste frequency into standalone measures of ability.
Put the guide to work
Rewrite one candidate note so every judgment points to a work sample. Add a follow-up wherever the evidence is incomplete.
Download the editable worksheet to record the decisions and open questions with your team. The examples in this guide are illustrative, not customer results or validated selection rules. Product availability and setup depend on the workflow agreed for your account.
A focused checklist, a worked scenario, and questions to resolve together.
Frequently asked questions
Can we see more than a total score?
Yes. Skill and section results, candidate work, and available session evidence provide a more detailed view of assessment performance.
Does slower typing indicate weaker technical ability?
No. Typing patterns are contextual session data. They should not be used as a standalone proxy for intelligence, competence, or integrity.
Is the highest-ranked candidate always the best hire?
No. A ranking reflects the assessment and its scoring. The final decision also needs the role's priorities, interview evidence, and a review of any gaps in the evaluation.
Keep building your hiring workflow
- Assessment integrity: A flag is a question. Find the evidence before the verdict.
- Interview copilot: They already took the test. Make the interview add something.
- Recruiting copilots: Give the agent a task. Give the exception a person.
Explore candidate analytics with LunaPrompts or browse all 19 hiring field guides.
