Anti-money laundering scrutiny is growing in the payments space

A hybrid approach combining rule-based models with AI may be more effective in reducing false positives and identifying suspicious behavior.

Money laundering is big business. The United Nations estimates that anywhere from $800 billion to $2 trillion is laundered globally each year. That equates to between 2% and 5% of global gross domestic product, roughly the same as the planet’s agriculture industry.

Regulators are focusing more anti-money laundering (AML) efforts on payment providers. Several new regulations have been published over the last few years, as the pandemic has boosted the volume of online transactions.

One of the key responsibilities for payment providers under AML regulations is to monitor transactions. There are a few key building blocks to consider when evaluating an effective transaction monitoring system, the first being data, which comes in two forms – live data from processing payments and static data from customer relationship management systems.

Sign up for the free BAI Banking Strategies newsletter and get industry insights delivered to your inbox.

By clicking the Subscribe button, you acknowledge that you have read our Privacy Policy and Terms of Use and agree to be bound by them.

The second is a detection engine. Data feeds into the detection engine, which applies rules to transactions to determine whether they pose a risk. Most financial institutions automate this screening process to cut costs, save time and remove friction from the customer journey. Technologies like artificial intelligence are helping to accelerate it even further.

The third building block is the user interface. When a transaction comes up as suspicious, the interface sends an alert prompting the compliance team to investigate it. At this point, human input may be required to review the transaction and the customer’s track record.

The biggest challenge for payment providers is finding a transaction monitoring solution that fits into their infrastructure quickly and without creating friction. Such solutions must reflect an understanding of the IT infrastructure used by payment providers and the unique issues they face. A further consideration is how the solution is deployed. Cloud technology is typically leveraged because it simplifies the integration process and it’s scalable, so users don’t have to invest in additional hardware to meet growing demand.

Another challenge for payment providers is the high frequency of false positive alerts. The rigid and rules-based nature of these solutions, many built on legacy infrastructure, means that 95% of alerts turn out to be false positives. Manual counterchecking adds friction to the customer journey and increases costs for FIs.

Technology can help by further automating the process of monitoring transactions. That starts at the integration stage. The use of flexible APIs allows providers to seamlessly integrate with their selected platform without adding complexity to the payment infrastructure.

And rather than a standard rules-based approach, superior AML compliance platforms will take a hybrid approach whereby the platform combines rule-based models with AI to monitor transactions (in real time, where necessary). This approach has become more effective in reducing false positives and identifying suspicious behavior more accurately than its traditional counterpart without AI integration.

Transaction data gathered from each client – either on an ongoing basis or from data gathered over the preceding 12 months – enables the AI algorithm learns the patterns and behaviors associated with legitimate transactions to distinguish them from suspicious transaction activity.

Over time, the algorithm learns how to replicate the client’s human decision-making process with increasing precision and automates the process moving forward, ultimately allowing banks, payments companies, and financial institutions to operate more safely, efficiently and strategically.

Tobias Schweiger is CEO at Hawk AI.