Anti Money Laundering

An innovative business intelligence application that will have to interface with the data stored in bank IT systems: both data concerning the handling of customers and transactions as well as data stored in the Archivio Unico Informatico (the centralised computer archive that is obligatory for banks and of the same importance, as per anti-money laundering legislation).

Using the most innovative data-mining technology, the system will not stop at detecting cases for investigation, but it will also forecast the risk of possible illegal actions of clients that are not among those cases reported for investigation. More specifically, the system will use the concept of automatic learning (machine learning) whereby it will be able to learn previously unknown causal relationships between the statistical distribution of data and the possible risks connected to money laundering, as well as take advantage of the relative heuristic algorithms and artificial intelligence, mainly based on the use of neural networks and Support Vector Machines (SVM).

In terms of the data that will be handled, the system will use data on relationships and transactions, accounts, funds and occasional transactions, reporting problems and analogies concerning a customer’s identity, the actual account holder (in the case of companies), companies owned by trust companies or foreign companies, identifying those transactions that could be considered split operations.

The project’s concept stems from an actual shortfall in the bank industry, where checks on customers are done manually, without the help of a specialised system and are only geared towards detecting problems, rather than preventing them.