Dedicated Algorithms for Risk Minimization in Suspicious Financial Transactions

Under EU Directive 2015/849 of the European Parliament and of the Council on the prevention of the use of the financial system for the purpose of money laundering or terrorist financing, it is necessary to identify both individuals and transactions of a certain degree of risk. The process of identifying the risk of both customers and transactions considered suspicious lies at the root of systems aimed at preventing money laundering and terrorist financing. Such systems are called AML (Anti-Money Laundering) systems. An important step in calculating a client's risk is to check his / her existence in lists of suspicious or possibly suspicious persons, also called sanction lists. Classic search methods involve large processing capabilities. Taking into account the obligation of all financial institutions to implement these methods, there is a need to implement a fast and secure search flow. Therefore, the attention was drawn to the searching techniques for artificial intelligence. These kinds of methods include advanced machine learning for optimizing the whole searching process: the system is able to detect certain patterns and identify new ones based on certain characteristics of the search query and by identifying similarities between words.