Banks must speed up fraud interdiction to keep pace with real-time payments

Brian Keefe, senior pre-sales consultant for NICE Actimize and a former federal financial crimes investigator, is confident that mid-market banks and credit unions are accelerating the robustness of their fraud-interdiction efforts in 2024 to match the speed of real-time payments.

Keefe told BAI he believes the redoubled determination of financial services organizations to mitigate fraud stems from the sting of their past fraud losses as well as their reevaluation of current procedures to ensure compliance with rapidly evolving regulatory requirements.

Banks that use a multilayered approach melding the best of automation with human expertise will become more efficient at detecting and preventing fraud, he says.

What are the leading current events driving research and investments in AI-powered fraud solutions?

There’s been a great deal of negative news about financial institutions, whether it was the collapse of several banks in 2023 or the growing fraud that impacts them. Banks are dealing with significant risk, including cyberattacks and a growing variety of fraudulent schemes. They’re responding by shifting their attention toward leading AI-powered solutions to ensure they’re up to speed with compliance. They’re adopting these new technologies to help match the speed of real-time payments. Globally, financial fraud is projected to reach $41 billion per year by 2027, according to ACI Worldwide. With the recent introduction of the FedNow instant payment service, the urgency to explore advanced fraud-fighting solutions is greater than ever.

Are these emerging fraud threats increasing regulatory scrutiny?

Yes, because regulators are increasingly looking at the real-time products and services that banks are delivering. Regulatory agencies want to ensure that these real-time payments systems that banks are adopting are compliant with new or existing regulatory guidelines. Regulation E, which protects consumers when they use electronic fund transfers, is continually changing. Regulators are requiring banks to shoulder greater liability when a customer is defrauded. FedNow mandates that participating financial services organizations undergo certification to ensure that they’re ready to combat instant-payment frauds that are growing at an alarming rate.

How will these emerging threats impact small and mid-size banks that lack the resources of larger institutions?

Fraud can have a disproportionate impact on smaller institutions because they can’t absorb the losses that would be much less consequential for larger banks. That impact can trickle down to a loss of customer/member confidence and even a loss in the customer/member base. Fraud losses can inhibit a smaller bank’s ability to expand its footprint, to offer new products and to adopt new technologies that can help them stay ahead of real-time fraud.

What’s an example of a new technology to prevent and detect fraud?

Generative AI (Gen AI), such as Chat GPT, is the most significant. This new technology can help banks and credit unions to review possible fraud more effectively and more efficiently. It can alert fraud mitigation teams to suspicious activities and patterns that go beyond normal boundaries. Gen AI and machine learning allow banks to free up valuable time and resources normally devoted to detection, prevention and investigation. It can also enhance accuracy by reducing false positives. Purpose-built Gen AI that’s customized to a bank or credit union’s specific needs can add many other efficiencies as well as provide analysts with critical information. With natural language processing, Gen AI understands the questions posed by fraud investigators and responds appropriately.

Does this technology provide gains in efficiency and inaccuracy?

Yes it does, because the traditional fraud alert review process is very labor intensive. It’s manual and linear in nature. By that I mean an analyst or an investigator spends probably 80% of their time manually reviewing and triaging alerts, whether false positives or negatives. The analyst must prioritize the level of an alert, determine a possible solution and create a case. For example, an institution that reviews 6,000 fraud alerts a month might require close to 20 analysts. Generative AI can assist them in many ways. In addition to reducing false positives, it can triage and de-prioritize alerts as well as produce patterns of detection that would help analysts detect future threats. Gen AI significantly reduces the non-investigative work and increases value. Adopting these advanced technologies becomes more important because the volume of real-time transactions is projected to increase as more and more institutions begin to offer real-time payment options.

Brian Keefe is Senior Pre-Sales Consultant for NICE Actimize.