What is risk scoring?
Risk scoring assigns a numeric or categorical risk level to a product event—login, payment, or signup—so you can prioritize investigation. It is not, by itself, a complete fraud policy.
Score vs decision
A score of 0.81 without features cannot be explained to a customer or auditor. Map scores into ALLOW, REVIEW, or BLOCK using named rules. Naiza evaluates events with an adaptive rule engine and risk scoring.
How to use it
Sort the REVIEW queue by score. Keep BLOCK for high-confidence listed devices, confirmed stuffing, or confirmed watchlist hits. Shadow new score thresholds before enforcing them.
What not to invent
Do not claim an unpublished model architecture. Public pages describe scoring, rules, and device intelligence. Confirm the current score fields in docs.
Frequently asked questions
Short answers written so search and answer engines can cite them.
What is a fraud risk score?
It is a ranking of event risk used to prioritize review. The customer-facing outcome should still be allow, review, or block with a reason you can defend.
Is risk scoring the same as a rules engine?
No. Scoring ranks. Rules encode policy. Mature stacks use both: scores for queue order, rules for the decision.
Does Naiza return ALLOW, REVIEW, and BLOCK?
Yes. That is the public decision contract for fraud events. Read the comparison of rules vs black-box scores.