American Express’ credit and fraud models completely use machine learning and AI. 

“In fact, the first time American Express used machine learning in the company was back in 2010 in our fraud models,” Anna Marrs, group president of global commercial services and credit and fraud risk at the financial giant, said Sept. 4 during the Scotiabank Financials Summit 2024.  

American Express leans on analytics and decision sciences within its fraud business to constantly tune and improve models, she said. 

“I think it’s an under-covered or under-understood part of American Express, just how much analytics and decision sciences are at the core of the enterprise,” she added. 

Amex is investing more in decision science resources, including its decision science resource centers, she said. 

Its latest resource center, the Singapore Decision Science Center of Excellence, was announced in November 2023.  

It focuses on researching and developing AI and generative AI for servicing, risk and technology, according to the company. 

“Fraud is a growth industry and it’s incredibly innovative, and generative AI is a tool for big companies to improve security or productivity,” she said.

It’s also a tool for fraudsters, which is why Amex works to stay ahead, she said. 

Register for the upcoming complimentary webinar presented by Bank Automation News: “The future of open banking: Payments meet data,” on Tuesday, Sept. 17 at 11 a.m. ET. Register for the webinar here.