Many financial institutions are developing and implementing Artificial Intelligence (AI) models with the goal of making their money laundering controls more effective and efficient. Crucial to the success of these advanced solutions, as well as basic Financial Economic Crime (FEC) processes, is having the data base in place. This has been and continues to be a major challenge for many financial institutions such as banks.
This was one of the key insights from the FEC Data event which recently took place at the Benelux headquarters of financial services service provider Delta Capita. During this event, representatives from various banks and fintechs came together to talk and learn from each other on the topic of FEC Data.
"We wanted to host this event so that participants could spar with each other about the challenges and solutions they are facing and thinking about within their own organizations," said Rob van der Kruijs, who is closely involved in Delta Capita's Data & Tech practice following his move from ING.
There is much belief in the potential of (Generative) AI and so during the event, great examples of successful implementations passed by. For example, smart assessment models aimed at reducing false positives have been implemented in the Transaction Monitoring (TM) domain. And within the Customer Due Dilligence (CDD) domain, for example, they are looking at how Gen AI can help with (more) automatic generation of summaries of customer inquiry interviews.
But while there have been initial successes with these advanced capabilities, participants also recognized significant challenges. For example, when using an AI-based Transaction Monitoring model, rather than a simpler "rule-based" TM model, it is important to pay more attention to how an outcome was created.
AI models are often considered a "black-box. Therefore, for the analyst to create confidence in the results of the model, it is crucial for the user to understand the underlying decisions made in the model and how they led to a particular outcome.
In addition to the technological developments that many banks focus on, there are simultaneously trends & strategic pillars within these organizations that call for the data base to be in order. For example, positive customer experience is seen as a key focus for many banks.
The contact moment between the customer and the CDD analyst during a customer inquiry, therefore, can play an important role in a positive or negative customer experience. It is therefore essential for banks to have its CDD customer information accurate and up to date so that when this customer contact occurs, it is as smooth as possible.
In addition, if a bank wants to work risk-based, it is important that the data from the various FEC domains is readily available and can be combined. With these insights, the most complete possible customer profile can then be formed on which an analyst can then make his/her analyses.
However, participants at the event mentioned that this very basis, having good quality data from multiple sources, is a challenge in practice. Because different systems are used and the FEC domains have limited cooperation with each other, it is often difficult to obtain the required data in a simple way.
However, this hygiene factor will have to be met if banks really want to organize their money laundering controls more efficiently and effectively. It is therefore necessary that banks pay extra attention to this by means of thorough data management, among other things.
Banks will also need to remain firmly committed to advanced technologies in combating Financial Economic Crime. However, in this context, the basic data will also have to be in order for the potential of AI models to fully mature within the FEC domain. After all, this data is the main input for these models and will therefore ultimately determine how successful AI is in performing money laundering controls more effectively and efficiently.
