Walmart experiments with A.I. to monitor stores in real time

Who's minding the store? In the not-too-distant future it could be cameras and sensors that can tell almost instantly when bruised bananas need to be swapped for fresh ones and more cash registers need to open before lines get too long.

Walmart, which faces fierce competition from Amazon and other online retailers, is experimenting with digitizing its physical stores to manage them more efficiently, keep costs under control and make the shopping experience more pleasant. On Thursday, the retail giant officially opened its Intelligent Retail Lab inside a 50,000-square-foot Neighborhood Market grocery store in Levittown.

Thousands of cameras suspended from the ceiling, combined with other technology like sensors on shelves, will monitor the store in real time so workers can quickly replenish products or fix other problems.

The technology will also be able to spot spills, track when shelves need to be restocked and know when shopping carts are running low. Cameras, for example, can determine how ripe bananas are, based on their color, and workers will get an alert on their phone if they need to be replaced.

One reason the discount retailer chose Levittown for the retail lab is that the store is one of the busiest Walmart Neighborhood Markets in the country, said Ravi Jariwala, spokesman for Bentonville, Arkansas-based Walmart Stores Inc.  The retailer has 698 of them nationwide.

“Because of that volume, it gives us an incredibly rich amount of information,” he said.

Levittown also was chosen because of its proximity to Walmart’s engineering and technology teams, which are based in Hoboken, New Jersey. Those teams spent 18 months visiting the Long Island store, preparing for the launch, Jariwala said.

The Levittown store, the only Neighborhood Market among Walmart’s 12 stores on Long Island, opened July 2013 and has 130 employees and offers30,000 products.

The retailer, which has 5,355 stores nationwide, has no plans to roll out the lab concept to its other stores, Jariwala said.

Instead it will use the insight gained from the Levittown store to improve efficiencies chainwide. “I think what we’ll see instead is a continued evolution,” he said.

Walmart's deep dive into artificial intelligence in its physical store comes as Amazon raised the stakes in the grocery business with its purchase nearly two years ago of Whole Foods Market.

That's put more pressure on Walmart and other traditional retailers to pour money into technology in their stores. At the same time, they're trying to keep food prices down and manage expenses. Amazon has been rolling out cashier-less Amazon Go stores, which have shelf sensors that track the 1,000 products on their shelves.

Walmart's online U.S. sales are still a fraction of Amazon's online global merchandise empire, which reached $122.98 billion last year. But Walmart says more than 140 million U.S. shoppers visit a store in person or online per week, creating a treasure trove of data. In its latest fiscal year ended Jan. 31, Walmart generated more than $500 billion in overall sales globally.

"We really like to think of this store as an artificial intelligence factory, a place where we are building these products, experiences, where we are testing and learning," said Mike Hanrahan, CEO of Walmart's Intelligent Retail Lab and co-founder of Jet.com, purchased by Walmart three years ago.

Hanrahan says the cameras are programmed to focus primarily on the products and shelves right now. Sensors embedded in shelves will give the store extra information because they know what's at the back of the shelves that the cameras can't see.

Thed cameras do not recognize faces, determine the ethnicity of a person picking up a product or track the movement of shoppers, he says.

Some other companies have recently started experimenting with store shelf cameras that try to guess shoppers' ages, genders and moods.

There are signs throughout the Neighborhood Market, educating shoppers about how it is being used as a lab. Still, the cameras could raise privacy concerns.

"Machine learning fundamentally finds and matches patterns," says Steven M. Bellovin, a computer science professor at Columbia University and a privacy expert, who hasn't seen the new Walmart A.I. Lab. But he says companies run into trouble when they start to match behavior to a specific customer.

Hanrahan says Walmart has made sure to protect shoppers' privacy and emphasized that there are no cameras at the pharmacy, in front of the restrooms or in employee breakrooms.

Hanrahan says the company is embracing the labs in stores to better understand the real ways that technology affects customers and workers. It also wants to educate shoppers. Walmart has made it a point to not hide the technology, and small educational kiosks about it are set up throughout the Neighborhood Market. Shoppers can peer through a glass-enclosed data center at the back of the store that houses nine cooling towers, 100 servers and other computer equipment to process all the data.

Despite the signs and visible cameras, many shoppers, including Marcy Seinberg of Wantagh, didn't seem to notice or care.

"I am not bothered by it," Seinberg said. "If technology saves me money, I would be interested."

With Tory N. Parrish