Anti-Money Laundering (AML) is a significant burden for large and small banks, irrespective of their line of business. Money laundering techniques are evolving faster than we could think. In order to detect sophisticated organised crime, there is a compelling need to bring to light concealed connections among multiple but seemingly unrelated human money laundering networks, ties among actors of those schemes, and amounts of funds transferred among those entities. However, legacy systems with static rules and traditional approaches like the configuration of new rules and thresholds are ineffective as they generate ultra-high false positives and fail to identify false negatives buried deep inside the mountain of legitimate transactions. Result: Suspicious Activity Reports (SARs) are filed late or worse no filing at all.
Graph analytics also called network analysis, a modern technology that analyses relationships between entities, can be used to address this growing problem. This paper details Tookitaki’s Network Science module which helps its Anti-Money Laundering (AMLS) solution to automatically and accurately detect complex money laundering patterns. In line with Tookitaki’s overall efforts to use machine learning and big data analytics for AML purposes, the module was developed with ways to avoid subjectivity and ensure scalability. At Tookitaki, we envision that this advanced data analytics approach will enable financial institutions to capture complex money laundering transactions and stop the bad actors with high accuracy and speed, improving returns and risk coverage.
Tookitaki’s concept of Automated AML Pattern Detection Using Network Science was awarded the Monetary Authority of Singapore’s Financial Sector Technology and Innovation (FSTI) Proof of Concept (POC) grant on 17th December 2020. The FSTI POC grant provides funding support for experimentation, development and dissemination of nascent innovative technologies in the financial services sector.
