Criteria and weights, set up front
Before the first candidate applies, you define what matters and how much each thing weighs. The rubric is a first-class object bound to the role — not a note someone types after the interview.
Weighted scorecards
The math is the product.
Most ATSs record a gut-feel “Strong Yes.” FairHire records numbers with weights, and rolls them into one composite you can read, defend, and publish. A scorecard here is a first-class object — not a comment field.
What makes it fair
A rubric you can point to
One transparent composite
Each criterion contributes weight × score. The composite is the weight-normalised sum — Σ(weight × score) ÷ Σweight — always on a 0–10 scale. No hidden rounding, no mystery ranking.
The same number, everywhere
The composite on the pipeline card is the composite on the candidate page is the number your agent reads over MCP is the figure on the published fairness page. It’s computed once, by shared code, and never recomputed.
Applied equally to everyone
The same rubric runs against every candidate for the role. That’s what makes a comparison defensible: not that everyone scored well, but that everyone was measured the same way.
The one formula
No black box. Just a weighted average.
The composite is deliberately simple, deliberately readable, and computed by one shared function so it can never diverge between a recruiter’s screen, an agent’s tool call, and a candidate’s transparency page.
Give every candidate the same measure.
Set the weights once, score on evidence, and let one honest number carry the decision.