EMPLOY NETWORK ANALYTICS IN MACHINE LEARNING SYSTEMS TO UNCOVER PREVIOUSLY UNDETECTED MONEY LAUNDERING RELATIONSHIPS

If we are to use a military analogy, money launderers engage in the financial crime versions of Asymmetric Warfare; they operate in unconventional ways, specifically to confuse the conventional systems that you employ to identify and interdict them. I know that from a decade of personal experience in that dark art, and two subsequent decades of trying to catch operators with the same goals and objectives as I had at that time. If you are to be successful in AML, you need to be unconventional as well, and there are new tools that enable you to do so.

Compliance officers need to understand that money launderers stay up night and weekends, brainstorming just to beat you. They are constantly seeking to adapt and modify known tactics, so as to make them unintelligible to your inquiries, especially if you are using rules-based systems to  detect their operations. In plain language, they endeavor to operate using unconnected and therefore quite undetectable, systems and methods, betting that your traditional software's approach will be unable to ferret out any connection.  

Enter network analytics, which, if employed within an AI platform with machine learning capability, is capable of making connections among what it finds to be related entities, out of previously undetected relationships. Once such a relationship is located,  additional data is gathered on the subjects the program located, and supplemental analysis uses it to search for more links and relationships, adding them to the previously-extracted information to form a web or pattern.

The result can be the discovery of a hitherto previously unknown laundering operation that cleverly utilized disjointed and seemingly unconnected transfers, which were ultimately amassed elsewhere, outside of the scope of your inquiry. Network analytics, operating in an AI-powered system employing machine learning,  however, was eventually able to cobble together these totally unrelated transactions, and display them, exposing a covert money laundering pipeline that was previously fooling your old rules-based program. It is successful because it not only operates outside the bounds of what was a limited system of inquiry, it has the ability to move past the initial data recovered to find that hidden network, due to machine learning capability in an AI-driven environment.  

For further reading;

Network Analytics + AI = The Future of AML