Are the old ways of transaction monitoring dead?

Financial institutions continue to face the challenge of demonstrating a comprehensive anti-money laundering (AML) transaction monitoring programme that is designed to detect, and aligns with, relevant Federal Financial Institutions Examinations Council (FFIEC) red flags without explicit, consistent confirmation on whether the escalations (ie Suspicious Activity Report [SAR] filings) are correct or valuable. Historically, this led most banks to adopt the use of typology-based if/then rules, resulting in a significant volume of alerts to be reviewed and dispositioned, with only a small portion being identified as potentially suspicious. While machine learning models are touted as an obvious fix to this problem, many banks may find such solutions to be far too expensive, complex and/or resource intensive. In order to answer the question, ‘are the old ways of transaction monitoring dead?’, this paper offers and evaluates various practical solutions, ranging from simple to sophisticated, to reduce false positive alerts generated by traditional AML transaction monitoring applications.