Regtech company ComplyAdvantage has teamed up with anti-fraud fintech Resistant AI on a solution that alerts financial institutions to patterns that signal money laundering and other fraudulent activity.

The artificial intelligence (AI)- and machine learning (ML)-powered solution, which has been in the works since ComplyAdvantage was founded in 2014, offers banks two options for implementation, Arshi Singh, head of product at ComplyAdvantage, told Bank Automation News.

“[Banks] could either [implement] through API or they could just use our [user interface] and do a file upload,” Singh said. “We then do our analysis of that data using certain scenarios that [banks] would have pre-configured in our system, or using artificial intelligence, and we return in response notifications on anything that [is] suspicious.”

The technology runs on a “layer of AI on top of traditional rule-based or pre-configured risk scenarios,” Singh said, and triggers an alert when fraudulent activity is suspected. She listed examples such as a banking customer making payments to an entity impersonating the IRS or scammers attempting to take advantage of elders for financial gain.

The AI-based tool works by:

• Detecting unknown patterns of behavior and prioritizing risks from high to low;
• Identity clustering when fake accounts may be linked through contact information; and
• Discovering new risk scenarios, such as a person’s account being hijacked when payments are made to suspicious recipients.

“What we do is go beyond the rules and the prefixed scenarios that the customers may have and use these machine learning models and the artificial intelligence to identify new, previously undetected risks,” Singh told BAN.

Helsinki-based digital banking service Holvi is among the first users of the solution, according to a release.

Financial institutions lose between $800 million and $2 trillion in fraudulent transactions each year, with just 10% of those funds being discovered by banks, Singh noted.