Can A.I. Outwit Your Buying Habits?

It’s inflation season. The economy is rocking a 5.3% inflation rate, the highest it has been in more than three decades. At the same time, consumers are sitting atop extra savings ($3.7 trillion in America), itching with pent-up demand to spend regardless of price. If you’re on the corporate side of this equation, you can add massive shipping delays plus supply chain cluster jams and a crude oil chart that looks like a terrifying roller coaster. Got all that? You don’t got all that, because no one does.

Confused and frustrated, a growing number of companies are facing all this pricing mayhem by engaging artificial intelligence more. Machine learning comes with many plus sides: by anticipating what consumers will tolerate, firms can maximize pricing income. But it also tests consumer loyalty when prices go too high. “Brand loyalty is declining, and at the same time, consumers are becoming more accustomed to comparing prices between retailers, even when shopping in-store,” says Deepali Vyas, global head of Korn Ferry’s Fintech, Payments, and Crypto practice, adding that price perception is “crucial right now.”

Most firms are pursuing highly curated pricing through consumer apps. It works like this: you’re walking down an aisle in the grocery store with your phone in your pocket, and a coupon for your favorite chips pops up for you on the store’s app. The app is constantly learning your favorite brands and craving schedules, and develops a pricing strategy of what to put on sale or not. But with pricing AI still in its infancy, experts say such algorithm tracking is far from perfect. “Right now the brightest minds in retail can’t predict what Christmas sales are going to look like,” says Craig Rowley, a senior client partner at Korn Ferry. “So AI won’t either. The variables are too hard.”

He adds that “machine learning is not something you do once. It takes hundreds of times of looking at a problem and solving it, and every time you get a little better.”

In-house, AI pricing tends to work with the least human oversight when deployed within narrow silos. For example, rideshare and airline companies have long used congestion pricing algorithms, which spike prices during peak travel times. These algorithms are narrow and crunch a reasonably small set of variables. But if an airline or rideshare company wants to market a credit card only to its credit-worthy customers? “A human has to step back and say holistically we have this overall business objective, and we need all of these silos to work together to have this true north,” says Vyas.

Experts say a key element to making all this work is the talent handling it, and that the team overseeing such analytical machinations is more important than the technology. Many firms are opting for chief analytics officers, who are expected to ask cross-silo questions such as “Why are we marketing on both coasts when 80% of our sales are coming from middle America?” It is a challenging role. “The problem is that the quality of the data may not be as strong,” Vyas says. Indeed, MBA programs have recently emerged to help produce people with the requisite technical, data, analytic, and business skills. “You actually need a real data person to figure out the quality of the data before you can analyze that data set,” says Vyas.