The Friskore™ score runs 0–100, higher means cleaner. Five bands drive the decision: Verified Clean (85–100) and Low Risk (70–84) always pass. Moderate Risk (50–69) triggers FaceGraph™ — a verified face upgrades to pass. High Risk (25–49) blocks unless FaceGraph™ verifies identity and identity clusters dominate. Critical (0–24) always blocks regardless of any verification.

FaceGraph™ activates for Moderate Risk always, and for High Risk only when identity signals dominate the score. If behaviour signals — automation patterns, routing anomalies, or timing irregularities — account for more than 40% of the High Risk score, the respondent is blocked at Stage 1 without triggering FaceGraph™. There is nothing verification can resolve. For identity-dominant High Risk scores (VPN, geo mismatch, device, duplicates), FaceGraph™ fires and a verified face upgrades to pass.

All 8 clusters run in parallel in under 400ms. The respondent experiences no perceptible delay. When the score lands in the Moderate or High Risk band, a webcam prompt appears. If no webcam is available, the configured no-webcam policy applies — default is pass for Moderate, block for High. Blocked respondents are silently redirected to a screen-out URL and receive only a standard disqualification message. Every decision and its full reasoning chain is logged to the study dashboard.