Using AI to Fight Fraud and Money Laundering

Financial crime is on the rise. Money laundering and financial fraud are topics that are affecting financial institutions (FIs) around the world in many ways. According to a Nilson report released at the end of 2016, card fraud is expected to grow to more than $31 billion by 2020. In addition to the actual direct losses incurred due to financial fraud (such as credit card or check fraud), FIs are increasingly coming under the scrutiny of national regulators. Globally, FIs have been fined more than $321 billion since the 2008 financial crisis. In parallel to this, FIs are opening up more and more channels to keep up with the demands of both consumers and business customers, which enables new avenues for fraud.

How Can FIs Fight Fraud?

Traditionally, FIs have fought fraud using traditional rule-based systems. Our industry is flooded with products that are very difficult to implement, have a high false-positive ratio (i.e. the number of false alerts) and are very expensive to maintain. An official from a tier 1 bank once told me that it takes them more than 12 months to change rules in their fraud system. How can banks fight fraud effectively with such legacy tools? The famous German scientist Albert Einstein once said this about winning: “You have to learn the rules of the game. And then, you have to play better than anyone else.” But fraud has no rules! Fraudsters adapt their approaches more dynamically than FIs can keep up with. This is why traditional fraud detection solutions fail. They work only based on prior fraud experiences, i.e. “known fraud patterns”—and that’s no longer good enough for staying ahead in the fraud game.

On this note, I’m happy to announce a new addition to our Vynamic™ Security portfolio: Fraud Detection. In close partnership with one of the leading artificial intelligence (AI) fintech firms, Thetaray, Diebold Nixdorf’s latest solution employs AI to fight all kinds of financial fraud and provides a very modern approach to anti-money laundering. The patented machine-learning algorithms used are based on more than 30 years of research work and are able to deduce previously unknown fraud patterns by looking at highly complex Big Data sets in near real-time.

Unsupervised Machine Learning

By combining disparate sources of data, this advanced analytics platform deploys unsupervised machine learning to detect mathematical anomalies in the data set with very high accuracy. The intelligence in the platform goes beyond the basics by really trying to understand the relationship amongst the different data parameters without having any prior knowledge or rules. All this while keeping the lowest false positive rate in the industry. Combine this with Vynamic Transaction Engine, which offers industry-leading transaction processing capabilities, and FIs have a best-in-class solution that’s custom-made for the future.

As the market leader of the ATM industry, we haven’t forgotten our roots. Various sources of data, such as ATM transaction logs and CRM data, can be integrated into the platform to detect a wide range of fraud attempts such as ATM and account behavioral changes, as well as operational issues such as data integration problems. This goes beyond the ATM silos to get to a next-generation security architecture. And backing it all is our trusted global sales and support network, which ensures an efficient, streamlined integration process from beginning to end.

Our Fraud Detection solution offers robust protection on various fronts, from multi-channel fraud to anti-money laundering and ATM security, supported by our global network of tech experts. Interested in securing your business using future-focused solutions? Let’s start a conversation.