The art and science of fraud prevention

Inquisitiveness. Tenacity. A strategic mindset. Analytical thinking. An aptitude for technology. Problem-solving ability. Creativity under pressure.

Does your fraud-detection bot powered by artificial intelligence come with those characteristics?

BAI spoke recently with four thought leaders in the fraud detection and prevention field who agree that automation is invaluable—but human intuition is more essential than ever in the fight against fraud at financial services organizations.

Kevin Sanchez, senior vice president and head of U.S. Bank’s digital financial crimes team, says AI in fraud-risk management continues to evolve in ways that help human fraud teams get better at what they do.

Sanchez, who has led the team at the Minneapolis-based bank since 2020, looks for several qualities in hiring team members. “First, you need someone with a strategic mindset,” he says. “It’s critical that a fraud-risk manager is able to optimize fraud-risk management on one side of the coin and ensure that we are optimizing the customer experience on the other.”

And Sanchez likes candidates who think analytically. “There is a lot of problem-solving, a lot of troubleshooting,” he notes. “You need someone who can analyze data; who can draw some hypothesis or conclusions out of the data to solve challenges that will emerge and then fight the fire.”

Ash Khan, head of enterprise fraud management for BMO Financial Group, says it’s not just the members of his team who are responsible for preventing fraud at the Toronto-based bank. For example, BMO’s fraud foundational training program allows employees to visit the bank’s portal to watch videos on fraud-related topics.

Khan says most candidates for his team come from financial services, but we do hire outside people with strong data science skill sets. “We have also hired people from our call center or customer service reps from the branches. They understand customer interactions and how things work at the branch level, as well as the products and services we offer.”

He says fraud prevention is a balance of art and science. “We can’t rely solely on technology to solve our fraud problems. … The human element is critical. We need people with the right mindset and the right art to connect the dots.”

Canh Tran, co-founder and CEO of Chicago-based Rippleshot, which uses AI and machine learning to protect financial institutions from card fraud losses, naturally has a soft spot for technology.

“Technology can save users a lot of time through automation or analysis,” Tran says. “It can go through millions of transaction records and look at fraud and tell you what the trends are. That would take a human a very long time to do. That allows the manager to spend more time strategizing fraud-prevention measures.”

But, he adds, “Cultivating talent is just as important as the latest technology. You need both technology and talent. Technology can be a game changer and can solve a lot of problems. But it can’t be done alone. It still has to be decided by and operated by humans.”

Kevin Eack, who heads the Brannan Group, a Peoria, Ill.-based security consultancy, believes an effective bank fraud preventer “needs a curious and investigative mind that says: ‘Something doesn’t look right here.’ I come from a police background, so I’m tempted to say they need to think like a cop.”

Eack was a senior executive with the Illinois State Police and served two tours with the FBI, most recently as one of two fellows to then-FBI Director Robert Mueller specializing in counterterrorism. He also headed a financial crime team for State Farm Bank from 2011 to 2015.

Bots and other fraud-detection technologies are adept at identifying anomalies in accounts that might point to suspicious activity, he says.  “But those kinds of technology advances only take you part of the way. … Then the human takes over and studies what’s really going on with this behavior.  Is it fraud? Is it something worth bothering the customer about? Or could it just be a little change in pattern?”

As an example, he cites an analyst who detected a fraudster’s handiwork when an affluent couple’s large home equity line of credit was being rapidly drained. “Suddenly, $100,000, $150,000 and $200,000 is getting moved out of the account, all within a 24-hour period,” Eack says. “The analyst tried to call the customer to verify it was them moving the money.”

But the fraudsters had rerouted the couple’s phone because they knew the bank would call them to verify the movement of the money. The person who answered the couple’s rerouted phone number was an actor who assured the analyst that all was well and that the funds were being withdrawn to pay for a home-improvement project.

The fake customer, however, had an accent that didn’t match the customer’s profile, raising a red flag for the analyst, who went online and dug up some old information about the customer, including a different phone number. The analyst reached the wife, who said that neither she nor her husband had withdrawn the money.

“It’s that sort of digging, those extra steps, where the human makes a big difference because technology is only going to draw your attention to something that is different. It’s not going to fully investigate it.”

Edmund Lawler is a BAI contributing writer.

Find out where things stand with fraud protection and how it can be done more efficiently and effectively in the BAI Executive Report, “Finding an edge in fraud’s cat-and-mouse game”.