Mastercard Turns to Decision Intelligence AI to Stop Fraud

You’ve got to hand it to the credit card industry; it’s working hard to protect its customers and, by extension, its bottom line. Mastercard recently took things a step farther in terms of protecting customers from fraud by bringing out a complete artificial intelligence (AI)-based system known as Decision Intelligence. Its goal: to better identify a fraudulent purchase in the making and not stop users from making legitimate purchases by inadvertently jumping at shadows.

With Decision Intelligence in place, users will find a more reliable approval process waiting, and in turn, see a drop of false declines. Decision Intelligence represents the first such use of an AI used on a worldwide scale, and directly on the Mastercard network.

Decision Intelligence analyzes the standard day-to-day shopping patterns of a typical user and looks for anomalies, which can better suggest that a customer is buying something unusual. Then, the card company can step in and respond only to the anomalies, which improves the chances of not hitting someone with a fraud alert that never should have been.

Granted, this only goes so far; if the fraudster knows enough about his or her target to carry on as though it were the buyer, then the whole protection method is lost. That and the notion of actually using an AI as a protection measure on a credit card account may leave some potential users unnerved, and reconsidering their choice of credit card provider. That’s not likely to impact a very big portion of the market, of course, and the market that doesn’t want fraud protection to get in the way of a normal purchase is likely to be bigger than those afraid of Skynet rising up and shutting down a user’s access to eBay.

Mastercard’s new plan should prove valuable in the end, by solving a problem that some people are doubtless having in a fashion that will only terrify a handful of those people. That’s a net win, and given the credit card market as a whole right now, any protection method, even one that only works for a few, is welcome.