The same engine that powers screening and the Radar — evidence-based matching that shows exactly why a candidate scored the way they did, requirement by requirement.
Every score has a receipt. No more "the AI said so" — you can see exactly which requirement was and wasn't met.
e.g. "Node.js, 5+ years"
e.g. "6.2 years, 3 cited deployments"
A percentage grounded in that specific evidence
Feeds into the overall recommendation