Sardine Fraud Prevention is a platform designed to detect and prevent fraud for card issuing programs. The platform provides real-time risk scoring that evaluates transactions in milliseconds during authorization. It includes machine learning models trained on billions of sessions to identify high-risk patterns such as large foreign transactions, high velocity card spending, and geolocation mismatches. The platform addresses multiple fraud typologies including stolen cards, gift card fraud, friendly fraud, card cloning, card testing, and account linking. It uses device intelligence, behavior biometrics, identity verification, and consortium data to detect fraudulent activity. The system can identify stolen card usage by triangulating stated versus true device location with merchant location, and detect friendly fraud by analyzing device handling patterns and typing rhythm. Sardine includes pre-built rulesets with hundreds of rules built on a feature store of over 4,000 fraud detection features. The platform provides anomaly detection to identify irregularities such as spikes in transaction velocity or unusual Card-Not-Present activity. For compliance, the platform offers Customer Identification Program (CIP) capabilities including KYC, KYB, global watchlist screening, and document verification with support for Enhanced Due Diligence. It provides AML transaction monitoring with both real-time and batch processing, featuring pre-built rulesets with 500+ rules. Case management functionality allows teams to route cases for review, manage alert queues, automate risk decisions, and streamline activity reporting.