by Leo Jakobson | June 18, 2018
Using artificial intelligence (AI) to predict which rewards its credit card customers would be most interested in allowed HSBC bank to nearly triple the effectiveness of a customer rewards program. 

In a pilot program with 75,000 consumer loyalty program customers, HSBC worked with Maritz Motivation Solutions to use AI to predict which of four reward redemption options would most encourage customers to spend their points. Four emails were sent, offering either a travel, merchandise, gift card or cash award to customers. A control group received a random offer. 

Of those who redeemed points based on the marketing message, fully 70 percent redeemed for an award in the targeted category. Beyond that, the overall redemption rate was 40 percent higher for the AI-targeted group than for the control group.

"If it was completely random, then 25 percent of your redemptions would've come from each category," says Jesse Wolfersberger, senior director of decision sciences for Maritz Motivation Solutions. "This is nearly three times that, so that's pretty good. They received the perfect message which scratched an itch they weren't even sure they had yet."

From HSBC's perspective, that's a result that builds goodwill and cements customer loyalty. "By using AI to predict our consumers' redemption behavior, we've refined our loyalty program to 'speak' to our customers by promoting the perfect reward or experience for them," says Pablo Sanchez, head of retail banking and wealth management for HSBC in the U.S. and Canada.

While noting that the sample size was small by the six- to seven-figure size of mainstream consumer loyalty programs, Wolfersberger says this type of AI-directed program targeting is where the big data that companies have been collecting from their customers for years "becomes more actionable and more relevant."

He adds, "it's not just segmenting your audience by gender or age or region or something like that where there are these clean black-and-white lines that you could draw. This is more gray: because you exhibit these behaviors, we think that you're going to like travel. Or we think that you're going to like a gift card to Starbucks. As an industry, we're getting to a place that's going to completely revolutionize the customer experience and their loyalty experience." 

Indeed, this experiment only shows the tip of the iceberg for what AI can do, Wolfersberger says. What comes next is not just knowing that you want to redeem points for a merchandise award rather than cash or travel. 

"It's that we think that you like golf, so here's a golf club promotional email. In gift cards, we know that you like Starbucks, or we know that you like Home Depot," he says. "The other thing that's on the radar is making them more triggered in the sense that, instead of just sending all these emails out on a Monday and seeing what happens, maybe we find that you would rather have a coffee gift card on a Friday afternoon because that's when you need a cup of coffee."

Comparing it to the explosive growth of mobile smart phones over the past decade, Wolfersberger says he believes, "AI is going to be involved in everything we do -- our day to day lives are going to be dominated by algorithms that drive you to work as well as algorithms that will greet you at the coffee shop."