Artificial intelligence might be doing Chad Funk’s job in five years, the fraud analyst at BrooksRunning.com tells Digital Commerce 360.

“Machine learning and AI are very much the future of this industry,” Funk says. Funk is on a team of four employees onthe manual review team at the sports shoe and apparel manufacturer, where he approves or denies transactions that its fraud prevention provider has flagged as possible fraud.

But this role, he says, is likely to disappear in five to 10 years.

“Manual review will go away. There won’t be any need for it because AI can do it better and faster,” he says.

But right now, Funk is confident his team is still necessary to prevent fraud. BrooksRunning.com has a manual review rate of 1.2% of all transactions and a chargeback rate of 0.02%, he says. These numbers are “quite good,” in his view, and better than industry benchmarks that he’s seen: a manual revie rate of 5.2% and a chargeback rate of 0.09% for footwear apparel merchants in 2021.

The average chargeback rate for an ecommerce merchant is 0.49%, according to a July 2022 survey of 60,000  merchants from fraud prevention vendors Chargebacks911 and Card Not Present.

“For humans, if they are good at their jobs, humans can still outpace AI,” Funk says. “It’s not at a level it truly needs to be at to kick people out of a position,” he adds.

Many in the industry agree artificial intelligence can get an online retailer most, if not all, the way there in preventing fraudulent transactions. But humans are still needed in some capacity to fight fraud. For many, that’s a manual review team to further investigate a transaction that has certain attributes that have triggered the machine to flag it as risky. But others say with AI-based fraud prevention, manual review teams to review transactions will not be needed in the future, or even today.

The cost of fraud and managing fraud

Fraud is an expensive problem. A report from fraud prevention vendor Signifyd finds that for every $100 in fraud order losses,