Despite an estimated $110 billion global investment in anti-money laundering compliance, 2019 was a year filled with cases of banking fraud and money laundering. In the Netherlands, the country I reside in, het Financieele Dagblad reports that approximately €13Bn is laundered every year. This is a significant number for the small country, as it is almost 1.5% of GDP.

Last year, five major US institutions including the Financial Crimes Enforcement Network and the Federal Reserve System offered a joint statement on ‘innovative efforts to combat money laundering’. They urged a heavily regulated industry to be more open and – creatively – use new technologies to combat fraud. Such technologies are artificial intelligence (AI), machine learning (ML) and natural language processing (NLP). So going into 2020, what can banks do to ensure they understand all the data they have access to, and actually know their customers, in order to avoid fraud?

Data makes the world go ‘round
Although banks are already able to go through piles of data and spot fraudulent behavior, they can and should do it much better. The problem, at its core, lies in the vast amounts of data that they manage. Often, these data pools are siloed, each with their own legacy platforms or software to manage it. This makes it challenging to connect the dots and spot the irregularities. Banking fraud is a sophisticated affair so the goal is to go from being able to spot simple anomalies, such as customer’s card being used overseas or a large ATM withdrawal, to being able to pick up on increasingly complex patterns of illicit behavior.

The key to improvement is consolidation. Looking back at the statement of the US federal agencies, which encouraged the use of innovative approaches, this will go hand in hand with the use of AI, ML and NLP. Such technologies enable the efficient and thorough consolidation and analysis of all data – allowing increasingly complex anomalies to be spotted more efficiently and timely. The implementation can, for example, lead to fewer false positives, or false alarms, which happens when systems flag seemingly fraudulent behavior that is in fact appropriate. According to research by Accenture, these technologies have already enabled a global bank to drop its alert volumes by 20% without influencing its risk appetite. We experienced similar positive results during a consolidation project for a local bank.

Compliance and tech – friend or foe?
Contrary to what you might think, technologies such as AI also enable compliance. Take for example MiFID II, which requires all communication, whether written or verbal, to be recorded and filed for 5 years. With AI, banks can use cognitive voice-to-text solutions to transcribe and process these conversations more effectively. Furthermore, when it comes to GDPR, banks will also benefit from AI as personal details can easily be safely processed by automation tools, under supervision of qualified employees.

As with any technology, there are two sides to this coin. AI and ML have huge potential for banks to significantly minimize banking fraud through analysis. However, when it comes to the handling of big data, ethics becomes a big concern. It would be naïve to not acknowledge the current issues that need to be addressed. As algorithms are initially trained through validation and supervised learning by humans, it is inevitable that these algorithms will have some kind of bias built in. After all, the human supervisor can only teach and validate what he or she knows. But these downfalls don’t make these technologies any less valuable – it just means they require awareness and mindful, strategic implementation.

New year’s resolution
Properly implementing technologies such as AI, ML and NLP doesn’t come cheap. But considering the potential fines, loss of customers and reputational damages, these investments are more than worth it. I would therefore urge banks to make 2020 a year of transformation and embrace the analytical powers that are out there. Proper anti-money laundering compliance requires consolidation, agility and consequently customer insight. When done properly, it will inevitably lead to improved know your customer practices and better anti-fraud policies.