How AML compliance is intersecting with payment fraud prevention

Transaction fraud and money laundering are typically siloed areas within banks, but both require a keen eye to finding the right patterns.

Hawk:AI is working with its first U.S. client, North American Bancard Holdings, to provide real-time screening and monitoring to strengthen NAB's anti-money-laundering compliance and decrease false positives that require employees to investigate.

While AML compliance is at the forefront of Munich-based Hawk:AI's machine-learning offerings, it can also flag other fraud scenarios and incidents.

"Although the mandates and objectives of AML and fraud may still be unique, AML and fraud leaders are recognizing they cannot continue to fight crime in silos," said Charles Subrt, senior analyst in the AML and fraud practice at Aite Group. "The historical divisions are now becoming less important and less clear as the end goals are similar — it is about stopping crime and protecting financial organizations and their customers."

Tobias Schweiger, CEO and founder of Hawk:AI
"Sometimes there is overlapping between AML and fraud when screening, but the major difference is that when looking at fraud, you are trying to reject the transaction that is harmful ... but in AML it is not the financial damage as much; it is more of a compliance question," said Tobias Schweiger, CEO and founder of Hawk:AI.

Still, screening for AML activity and payment fraud attempts require looking for different triggers.

"Sometimes there is overlapping between AML and fraud when screening, but the major difference is that when looking at fraud, you are trying to reject the transaction that is harmful," said Tobias Schweiger, CEO and founder of Hawk:AI. "It might be a financial loss if that is executed, like someone not paying for a good that was not received, but in AML it is not the financial damage as much; it is more of a compliance question as there are rules and regulations across the world in moving money and what to look for in anti-money-laundering regulations."

The damage with AML is less about losing money, but more about losing reputation and running afoul of regulations, Schweiger added. "It's a different type of risk you are trying to manage by screening for AML vs. fraud."

NAB says it turned to Hawk:AI mainly because the company could tailor its machine-learning solution to existing payment systems and data sources, while also having the capability to deploy the solution on its cloud-first software stack.

“We believe Hawk:AI is unique in its ability to combine the right mix of methodologies to detect suspected money-laundering,” Jim Parkinson, chief experience officer at Troy, Mich.-based North American Bancard, said in a statement to the media. “Hawk:AI was also able to work most efficiently with our data science, software and cloud infrastructure teams, successfully delivering the project under time constraints. Their payment, compliance and cloud knowledge proved invaluable.”

In screening for money laundering, Hawk:AI's machine learning monitors for patterns in funds being sent to specific countries, particularly those commonly blacklisted as areas supporting terrorism or other crimes. Or it tries to catch the "money wheel" pattern in which multiple people take money from someone and forward almost the same amounts to someone else.

There are hundreds of deals like that unfolding and it takes an intelligent system to catch those patterns and try to connect the dots between many different accounts, Schweiger said.

"You are always looking for the unknown, so there is always a chance that you will pick up on something that is not necessarily money laundering, but just a fraud attempt," he added. "You pick up on those things and it makes it almost like a side product of screening for money laundering."

Because fraud and AML prevention and their delivery methods remain separate entities at most financial institutions, it is important that vendors have different sets of overlays, interfaces and outputs.

"Fraud prevention's job is to interdict and stop the bad guys from stealing money in real time," said Julie Conroy, research director and fraud expert at Aite Group. "Often, the regulatory expectation for the AML function is to detect the bad activity, report it to regulators, but not always interdict, so the cultures of fraud and AML tend to be quite different."

Still, financial institutions and companies are seeking baseline patterns of good behavior and to find anomalies, Conroy added.

"That is why we see vendors like Hawk:AI, Feedzai, Featurespace, FICO and many others targeting both competencies with their engines," she said. "I’d say with the merchant acquiring use case, these competencies are even more closely intertwined, since often a fraudulent merchant storefront is used for both fraud and money laundering purposes."

Hawk:AI has been serving clients in Europe for nearly four years and intends to concentrate on the U.S. market in the coming months after landing NAB as its first large customer.

"We have not made decisions to go more global in our presence, but in our past experience with payments, we realize it is becoming more global and money laundering is a global thing," Schweiger said. "The criminals don't stick to one country, so it really has us set up as a global business anyway."

It's more important, at this time, that Hawk:AI sets up its data centers within the country it is serving clients. It considers AML compliance a market in which all financial institutions and payment providers are upgrading their compliance to keep pace with a maturing prevention industry.

"A U.S. customer wants to be served from a U.S. data center, as they don't want the data leaving the country for something like this," Schweiger noted. "And it is the same in Europe."

Ultimately, Hawk:AI is part of a trend in which AML protection is undergoing increased innovation with providers and customers both spending more on the technology.

"There are many benefits with increased artificial intelligence adoption," Aite's Subrt said. "Advancements in computing power and new techniques such as entity resolution, robotic process automation, network and link analytics and others can process massive data sets and elevate financial crime monitoring."

It's an important scenario for financial institutions that are under increasing pressure to meet the letter and spirit of the AML law while seeking ways to detect potential problems and decrease false positives, Subrt added.

For reprint and licensing requests for this article, click here.
Fraud detection AML
MORE FROM AMERICAN BANKER