Amazon Stores CEO Says AI May Spawn New Retail Formats

Artificial intelligence (AI) is going to radically transform the retail sector and could result in the creation of new shopping channels, according to Doug Herrington, CEO of Worldwide Amazon Stores.

“We really haven’t had a technology revolution as large as this since the start of the internet,” Herrington told delegates at the National Retail Federation’s (NRF) Big Show.

He acknowledged that the shift to mobile and the rise of social media platforms were transformative, but both pale in comparison to the impact AI is going to have on the sector.

“It’s going to touch all of us. It’s going to lower costs, improve quality, help us develop new customer experiences, and it may even spawn new retail formats.”

How Amazon is using AI today

A couple of years ago, Herrington started an AI ‘show and tell’ so his teams could share ideas for leveraging the technology. He intended for this to go on for a few months, but the initiative is still running today, illustrating the ever-growing list of use cases that retailers are discovering for AI.

“[The teams] might be rebuilding a feature or a product that we already have but making it cheaper or better, or they’re inventing some totally new experience that felt like science fiction two years ago.”

Rufus, Amazon’s conversational shopping assistant, is a good example of this. “We’re up to half a billion [customer] questions that we just wouldn’t have been able to answer before in the search box, things like ‘tell me about protein powder’ or ‘can I use this fishing reel in saltwater’. It’s really opened up a whole new level of customer convenience.”

Herrington also cited customer reviews as an area that is ripe for AI disruption. Amazon has hundreds of thousands of customer reviews, and many of these are now neatly summarized thanks to generative AI. By surfacing this to the customer, Amazon is able to help them make quicker, more informed decisions.

AI is also helping Amazon with the perennial challenge of sizing and fit. This is fashion’s unsolved problem and Herrington believes AI can be part of the solution. “The biggest reason people might return [apparel and shoes] is fit issues,” he said. “We’ve got a team that has built a model, ingesting all of the size charts of different brands, all of the customer purchasing patterns, and all of the feedback that we get both from returns and customer reviews.”

Amazon is then able to produce a very specific fit recommendation beyond just size. For example, it might recommend a size large but highlight that the sleeves are probably going to be short on you.

Herrington shared that AI currently touches half of Amazon’s customer service interactions and it’s providing shoppers with “high-quality responses”. Amazon also now has a team that is rewriting product titles in real-time based on what shoppers are searching for.

AI is driving many efficiencies on the back-end as well, for example helping Amazon to identify fraud and stamping out fake reviews. Sellers can now quickly create new listings with just a few words or a photo, and there are countless use cases across the supply chain.

Erring on the side of innovation

Amazon has always been an innovator, but contrary to popular belief that’s not what sets them apart from their competitors, according to Herrington.

“A lot of companies say they love innovation, but they also really dislike failure. They end up, intentionally or not, with cultures that try to avoid failure. And if you get too good at eliminating failure, you’re probably walking away from a lot of great things.”

Herrington shared how founder Jeff Bezos’ ambition was to create a culture that generates bold ideas and allows for “thoughtful risk” in pursuit of those ideas. Bezos felt that too many ideas got blocked at other organizations because it only took one senior leader to say no. He wanted to reverse this at Amazon, so that it only required one senior person to agree to a project for it to move forward. This, in turn has fuelled innovation, giving employees the confidence to take risks and challenge the status quo.

Herrington shared how, in his earlier years at Amazon, he himself once pitched a new business opportunity to Bezos. Bezos, however, was not a fan.

“This meeting was not going well. He was challenging my assumptions, disagreeing with my models… We get to the end and he says: ‘You haven’t convinced me. I still disagree with you but I’m going to commit and let you make this bet because I want Amazon to be a place that has a bias towards an institutional yes.’”

This might sound counterintuitive, but Herrington says it has yielded a much wider funnel for the retailer. Bezos would remind him that, in the history of business, more value has been destroyed by companies that failed to try something new than by companies that try things but ultimately fail.

“That advice has always stuck with me, especially right now with this AI transformation that we’re going through. I think we’re going to err a bit on the side of innovation.”