Explainable fit score

Candidate fit scoring that explains why this person, not just who ranks first.

Kempian shows skills, experience, compliance, location, missing evidence, and a clear next step. Every consequential action still passes through a recruiter review gate.

AR

Alex Rivera

Senior Platform Engineer · Remote

87%
High Confidence

Why this confidence score?

Skills match
94
Experience
88
Compliance readiness
91
Location fit
82

Missing signal: Shift preference

Request confirmation before proceeding to outreach.

Try it yourself

What goes into a candidate fit score.

Four factor groups, each with its own weight and evidence. Click any group to see exactly what it evaluates and what missing evidence gets flagged before outreach.

AI Confidence breakdown

Click any factor to see what it evaluates

Composite

92%

High Confidence

Skills match

contributes 30% to composite

Direct alignment between candidate-declared and verified skills against role-required skills.

What this factor evaluates

  • Stated skills in candidate profile
  • Skills inferred from project descriptions
  • Years of practice per skill (when available)
  • Technology stack adjacency for missing exact matches
Every factor score is logged with the source signals that produced it. Auditable per GDPR Art.22.

The anatomy of an explainable fit score.

Every candidate score in Kempian includes the evidence recruiters need before approving outreach, shortlist movement, or submission.

1

Signal name

Identifies what is being evaluated: match quality, compliance risk, readiness, workforce fit.

2

Contributing factors

Each signal broken into named sub-factors with individual scores and documented source data.

3

Confidence band

High / Medium / Low with semantic colour. Never just a raw number without context.

4

"Why this?" access

Every score has visible reasoning. Expand any factor to see the source evidence.

5

Missing signals

Gaps are surfaced explicitly. Not hidden in a lower score without explanation.

6

Action path

System recommends the next step based on confidence state and missing data.

7

Human oversight

Recruiter approves, holds, or rejects. No candidate is contacted without human confirmation.

Design principles

What the fit score is, and is not.

A score without explanation is never acceptable

Factor groups are always named and visible

Missing signals are surfaced, not suppressed

Human review gate at every action step

"Bias-free AI" language anywhere on the platform

Fully autonomous hiring decisions

Score shown without action path

Guaranteed fairness or perfect match claims

Human oversight architecture

Four human gates, not one.

EU AI Act Annex III Article 14 requires meaningful human oversight on high-risk AI systems. Kempian enforces four gate points. No gate can be automatically bypassed.

Gate 1

Demand intake

Recruiter approves demand before sourcing begins. AI cannot self-trigger a search.

Gate 2

Shortlist review

Recruiter reviews scores and approves shortlist before any outreach.

Gate 3

Outreach approval

Recruiter explicitly approves contact with each candidate. No automated cold outreach.

Gate 4

Client submission

Final human sign-off before submitting candidate package to a client portal.

See candidate fit scoring in a live walkthrough.

Factor groups, confidence bands, missing-evidence alerts, and human review gates applied to roles in your industry.