Interview Scores You Can Actually Explain
A number on a candidate means nothing you can defend. Fintra scores each competency, shows the evidence it used, and gives candidates a recourse path - the standard regulators now expect.
What explainable scoring means here
A black-box hiring score is a legal and ethical liability. Fintra decomposes every interview result into the rubric dimensions it measured, attaches the candidate’s own words as evidence for each, and records who reviewed it - so a score can be explained to a hiring manager, a candidate, or an auditor.
- Per-dimension scores instead of one opaque number
- Quoted evidence from the transcript behind each dimension
- A candidate recourse path to contest or request review
- Human sign-off recorded before any adverse decision
Anatomy of a score
| Dimension | Score | Evidence basis |
|---|---|---|
| Problem solving | 82 | Walked through trade-offs unprompted |
| Communication | 75 | Clear, but buried the key point |
| Ownership | 54 | Deflected on the failure question |
| Role knowledge | 80 | Specific, current, and accurate |
| Overall | 73 | Weighted per role rubric |
Recourse and human-in-the-loop
What happens on a low or contested score
- 1
Nothing auto-rejects
A low score is a recommendation; no adverse action happens without a human.
- 2
Recruiter reviews evidence
The reviewer sees the quoted evidence and can agree, adjust, or re-interview.
- 3
Candidate can request review
A recourse path lets candidates flag issues (e.g., a technical glitch) for a human.
- 4
Decision is logged
The final human decision, reason, and reviewer are recorded for audit.
Built for the bias-audit era
Frequently asked questions
Can I show a candidate why they scored the way they did?
You can share the rubric dimensions and the reasoning without exposing other candidates’ data. Because each dimension has quoted evidence, feedback is concrete rather than a bare number, which improves candidate experience and supports recourse and fair-hiring obligations.
How does this help with a bias audit?
Structured, per-dimension scoring against a consistent rubric - plus a full audit trail of transcripts, question paths, and human decisions - is exactly the record an independent bias audit (as required by NYC Local Law 144) needs to analyze outcomes across groups.
Does a human always review the score?
A human reviews before any adverse action. Scores can be generated automatically, but advancing or rejecting a candidate is a logged human decision. That keeps a person accountable and satisfies the human-oversight expectations of high-risk AI hiring rules.
Can we adjust the rubric and weights?
Yes. You define the competencies and their weights per role, and you can tune them over time. Because scores are decomposed, changing a weight re-derives the overall score transparently rather than shifting an opaque model.
Stay in the loop
One practical finance briefing a week - new guides, checklists, and benchmarks.
Make every hiring score defensible
Start free, no card required. Score interviews with per-dimension evidence and a built-in recourse path.
Talk to us