Fraud & Transaction Security API
Detect payment-fraud patterns, transaction velocity anomalies, synthetic-identity inconsistencies, invoice-fraud signals, supplier bank-change risk, refund abuse, chargeback patterns, and multi-account clusters—using caller-supplied payments, invoices, and identity records with local explainable rules only.
Detect suspicious combinations of caller-supplied payment events (declines-then-success, instrument churn, billing/shipping mismatch). Anomalies are not confirmed fraud.
Identify abnormal transaction frequency and amount velocity from caller-supplied histories. Velocity spikes alone are never labeled confirmed fraud.
Identify inconsistencies in caller-supplied identity records (sparse profiles, age/postal mismatches, shared contacts). Personal data is masked in responses.
Detect potentially manipulated invoices from caller-supplied invoice exports (duplicate numbers, tax anomalies, missing PO). Not confirmed fraud.
Identify suspicious supplier bank-detail changes near invoice timing from caller-supplied records. Bank values are last4-masked.
Detect abnormal refund frequency and amount ratios from caller-supplied payment histories. Refund rate alone is not confirmed fraud.
Analyze recurring chargeback/dispute patterns from caller-supplied payment records. Elevated rates are indicators, not confirmed fraud.
Group potentially related accounts using deterministic overlap of caller-supplied emails, phones, devices, IPs, and payment-instrument last4 values. Linkage is not confirmed fraud.