Bias or impact evidence
Track whether the vendor can provide a current audit summary, testing method, limitation statement, and subgroup-performance notes.
Vendor risk
Score AI hiring vendors before they create evidence gaps.
AI features often arrive through ATS, interview, assessment, and workforce vendors. This scorecard helps teams see which vendors lack audit evidence, human oversight instructions, notice support, or change-notification commitments.
Scoring criteria
Evidence a buyer can request before rollout.
Human oversight
Require instructions that explain intended use, reviewer responsibilities, override paths, and misuse scenarios.
Model and feature updates
Record whether the contract or support docs require notice when AI features, models, or scoring logic materially change.
Related checks
Collect the inputs before assigning a score.
After scoring
Use the score to decide which evidence workflow needs monthly upkeep.
Map the score to an evidence path.
Use the Evidence Hub to decide whether a weak score should trigger inventory cleanup, vendor renewal evidence, pre-audit prep, or a monthly operating rhythm.
Open evidence hubDecide if the score can go stale.
Model releases, feature toggles, audit updates, and renewal answers can change the risk file. Compare monthly updates with a one-time checklist.
Compare monthly workflowReview the paid pack path.
Check checkout access, download behavior, and support expectations before moving the scorecard into the Pro operating file.
Review checkout access