Transaction monitoring

Transaction monitoring is the ongoing, largely automated screening of financial transactions — individually and in aggregate over time — to flag activity that may indicate fraud or money laundering. It looks for patterns such as many small transfers kept just under a reporting threshold, or a sudden change from a customer's normal behavior.

It combines rule-based scenarios, specific, explainable patterns defined by compliance teams, with statistical anomaly detection or machine-learning models that learn typical behavior per customer or segment and flag deviations from it. Outputs feed a case management queue where analysts investigate and disposition each alert.

Transaction monitoring differs from point-in-time fraud detection checks made at the moment of a transaction; it also looks across many transactions and longer windows, which is essential for spotting money-laundering typologies that no single transaction would reveal on its own. The central operational challenge is alert volume: thresholds set too loosely generate more alerts than analysts can review, while thresholds set too tightly risk missing genuine cases — a trade-off usually managed through risk scoring to prioritize which alerts get reviewed first.