AI Money Laundering Watchdog That ‘Senses’ A Financial Crime

By Sara Miller, NoCamels September 24, 2023 4 minutes

Thetaray's AI platform tackles the global money laundering that the UN says is worth $800 billion each year (Depositphotos)

As the digital world expands to every area of our lives, we have become used to automated banking – with our transactions carried out instantly and with just one click.

But behind the ease with which we move our money about is still a meticulous process of verification, known as compliance, ensuring that our transactions abide by the letter of the law and are innocent of financial crime.

International financial transfers are vetted to ensure they are not money laundering in a slow and costly process (Pexels)

Israeli company Thetaray has developed a new AI platform that “intuits” that the billions of dollars, euros and even shekels that are moved around the world every day are untouched by the money laundering carried out by drug cartels, terror groups and even human traffickers.

And unlike other companies that will inadvertently flag innocuous transactions, Thetaray maintains that SONAR almost always only singles out genuinely problematic requests.

According to the United Nations Office on Drugs and Crime (UNODC), money laundering usually follows three steps: placement, or moving money from a direct link to a crime; layering, or hiding the money trail to make illicit funds hard to track; and integration, or making the money appear to be from a legitimate source.

All of these steps rely on financial institutions failing to spot the origins of the money – and the statistics show that do often they slide through the cracks, however much time and manpower is devoted to thwarting such endeavors.

Global Operations

The UNODC estimates that between 2 and 5 percent of the global GDP – some $800 billion – is laundered every year. It points out, however, that it is hard to know the true sum due to the secretive nature of the process.

Banks and financial institutions monitor all financial transactions for suspicious activity (Pexels)

Thetaray’s SONAR is a cloud-based Software as a Service (Saas) platform that plows through all the data of a financial institution, raising a red flag about any transaction that could be illicit in nature.

It works by ignoring the painstaking, established “rules based” monitoring of transactions that raises multiple false positives, Thetaray CEO Peter Reynolds tells NoCamels, and instead seeks out unusual yet apparently harmless behavior that might otherwise slip through the net.

Reynolds says that this means that SONAR has 95 percent less false positives as compared to other transaction verification programs, requiring less time and effort on false alarms and making it easier to spot genuine money laundering.

“If you and I worked in compliance, and we got 95 percent less alerts to look at and all those alerts were very detection worthy, you have much less to look at and the ones that [SONAR] did give you to look at were financial crimes,” Reynolds explains.

Sign up for our free weekly newsletter

Money launderers are familiar with the way in which financial institutions track fiscal transactions (Pexels)

People trying to launder money are just as cognizant of the rules and the violations that sound an alert, Reynolds says, and they make every effort to escape notice.

He cites the example of someone trying to pass multiple transfers of small sums instead of two or three larger transactions. SONAR can spot these indiscernible yet nefarious activities, Reynolds says, due to what he calls its “market-leading, state-of-the-art AI.”

More Than A Feeling

With machine learning that uses a constant stream of previous transactions to educate the platform about the difference between innocent and corrupt activities, the AI has been taught to intuit when a certain action does not “feel” right.

“AI intuition is that ability to sniff something out and say: based on everything else we see, based on how other models look, it just doesn’t look like a normal transaction,” Reynolds explains.

He compares this intuition to the feeling you might get from an entertainment venue that looks innocuous from the outside, but whose location or even exterior façade triggers your sense that something is out of place. It looks perfectly acceptable, Reynolds says, but an innate instinct tells you that this is not quite right.

Illustrative: Thetaray’s artificial intelligence platform ignores the ‘noise’ in transaction verification (Unsplash)

Founded in 2013, the Hod Hasharon-based company raised $57 million in its most recent funding round, with backers including Israeli entrepreneur Erel Margalit’s JVP and the Jerusalem-based OurCrowd.

SONAR is used by dozens of financial institutions worldwide, among them Santander, Payoneer, ClearBank and MFS Africa, and is hosted on Microsoft’s cloud computing platform Azure.

According to Reynolds, SONAR’s ability to weed out just the genuinely problematic transactions and ignore the “noise” is financially beneficial to companies who can relocate resources that previously went on the crucial work of monitoring transactions in order to meet international regulatory standards.

“It’s almost like turning compliance into a profit center because you can start to process the transactions that have no risk,” he says. “The transactions that are bad, you absolutely stop.”