Why Walmart Is Emerging As an AI Powerhouse

Analysts have characterized the recent strength in the stock market as an AI rally, but flying under the Magnificent Seven’s radar is Walmart — a company so vast that it literally has its own weatherman.

And as it turns out, the retail juggernaut’s scale and reach are proving to be tremendous assets in the AI race.

That’s because most top AI companies — like OpenAI, Microsoft, Anthropic, or Meta — operate in a primarily virtual space, processing unfathomably complex rivers of information into more digital information. AI-adjacent companies like Nvidia, Intel, and Oracle focus on providing the physical infrastructure upon which the AI machines function. Then there are the companies that are using digital intelligence to deliver physical results through automation and augmented experiences, like Tesla and Amazon.

Walmart, by contrast, has a vast and complicated set of physical challenges to solve as the largest retailer in the US — and the world. Those include everything from cleaning up spills in the dairy aisle to stocking shelves.

“We move billions of items around every month, every year,” Walmart US CEO John Furner said Tuesday at the Fortune Brainstorm Tech conference. He said the company has been developing machine learning tools and other automation projects since around 2015.

Furner said that the company’s AI models and supply chain automation help plan inventory to arrive at the right aisle at the right time, for example. One technique involves creating “digital twins” of each facility to model the movement of merchandise through the system on its way to customers.

Furner also said store associates increasingly have an AI chatbot handy via their handheld devices to help them better set priorities and help customers.

“It’s a combination of people being powered by technology. There’s a lot of judgment to retail and decision-making. And we’re in a very dynamic industry,” he said. “We think this next phase of physical AI in combination with Gen AI is going to be really helpful.”

The company’s head of e-commerce, David Guggina, told the Goldman Sachs Communicopia and Tech conference last week how AI is helping his team run experiments and fulfill orders at an increasingly rapid rate.

Guggina said his team is now able to work at breathtaking speed behind the scenes, too.

“What took a data scientist days or weeks before can now be done in minutes,” he said.

Related stories

AI also helps ensure that each of the company’s 4,700 stores has the kinds of products best suited to their local markets, slashing delivery times to minutes after a customer places an order.

“We’ve just completed the third inning,” he said by way of the classic baseball game analogy. “So we’re still early with regard to our automation journey in the fulfillment network.”

These digital-to-physical uses of AI are also complemented by a myriad of “micro agents” that handle tasks like tracking local event calendars or monitoring inventory levels.

Walmart, of course, is still fine-tuning its AI approach, and there have been hiccups.

The proliferation of bespoke Walmart-made AI agents eventually started to confuse users, the company told the Wall Street Journal.

The company has rolled many of those micro agents into four “super agents” designed to assist shoppers, merchandisers, programmers, and third-party marketplace sellers.

Still, because Walmart’s 20,000-strong global tech team builds so many of these digital and physical solutions in-house, the company is emerging as an unexpected AI powerhouse.

The company snagged former Instacart exec Daniel Danker in July to accelerate its AI efforts.

It’s also deepening its partnership this month with OpenAI via a new training program for associates and enterprise access to ChatGPT tools for frontline Sam’s Club employees to help operate their warehouse stores more smoothly.

After all, while chatbots might sometimes hallucinate answers, there’s no faking a cold gallon of milk on your doorstep.