The term “artificial intelligence” may be relatively new to some industries. But compared with peers in other verticals, most large financial institutions are old hands at the use of AI or machine learning, particularly for weeding out potential fraud risk in credit lending.
However, as AI has become “smarter” and fraudsters are becoming more pervasive and wily, financial service institutions (FSIs) and their service providers are beginning to find new ways to utilize AI for tracking down and reducing potential fraud in various transactions and areas.
“The greater use of AI simply proves the point that our traditional defenses are failing to protect us against bad actors,” said Adrian Talapan, co-founder and CEO at Fee Navigator, a service provider that works with FSIs and their merchant-customers. “It is the behavioral model of user interactions that now rules across industries, and AI is really good at learning normal behaviors and discrepancies.”
Fee Navigator uses AI in its service, to instantly analyze merchant statements, for acquiring side banks and their business customers.
Other emerging financial technology companies (or fintechs) and cybersecurity units from more established financial vendors are boasting about the integration of AI and machine learning in their offerings. Cases in point: Kount, an e-commerce fraud mitigation platform now owned by Equifax, utilizes AI in weeding out potential fraud for its FSI customers at online account-opening and login, to making payments and reviewing disputed transactions.
Recent fintech startup Resistant AI, as its name would indicate, makes broad use of artificial intelligence in its FSI fraud-fighting tools, which are aimed at finding potential risk in credit scoring models, payment systems and onboarding new customers. (Incidentally, the two-year-old Prague-based Resistant AI late last month closed $16.6 million in Series A funding, signaling the investment community’s growing commitment to AI for fraud detection.) As of late October, the Czech Republic-based financial tech vendor had reportedly signed up more than 30 customers, including global banks, insurance companies and fintechs.
“Machine learning, a subfield of AI, has become increasingly valuable in the fraud prevention space,” said Daniel Holmes, director for solutions consulting at LexisNexis Risk Solutions. “Machine learning helps banks and other online businesses look at their historic data at scale to determine what the key risk signals are within fraudulent transactions. These learnings can then be reapplied for future fraud prevention endeavors.”
Indeed, since AI-based systems can “learn” from their experiences and mistakes, these tools can make a powerful and even game-changing impact in how effectively and quickly FSIs can pinpoint and stop various types of fraudulent behavior — even if there are subtle differences.
“For example, it helps banks to understand what is normal for their customer, then allows them to compare the current event to that historic baseline in real-time, flagging any potential anomalies for investigation,” Holmes said. Since this is all conducted by the system “quickly and passively... [it happens] without interruption to the user experience.”
More recently, banks have discovered that machine learning can help them understand their customers at an overall consumer level, Holmes continued, “not just as a cardholder or a digital banking user.
“Using machine learning to remove channel silos and look at your customer as a collective entity helps banks maximize the use of the vast data they hold on their customers,” Holmes added, "driving optimal value from these assets when it comes to keeping them safe from cybercriminals.”
