How Walmart is using A.I. to make shopping better for its millions of customers
Amid the recent flurry of excitement around the transformational potential of ChatGPT is the fact that companies have been using artificial intelligence across their businesses for years.
But few companies have the ability to gather the massive data sets that power AI quite like Walmart. There are roughly 4,700 Walmart stores and 600 Sam’s Clubs in the U.S. employing a combined 1.6 million workers — or associates as the company likes to call them.
Deploying artificial intelligence and machine learning in ways that improve both the customer and employee experience across such a massive environment is the focus of Walmart’s AI strategy, said Anshu Bhardwaj, senior vice president of tech strategy and commercialization at the retail giant.
At warehouse chain Sam’s Club, the membership model provides the company with huge amounts of data about what members are searching for, what they like, what they’re buying and when. “[Customers] are generating all these breadcrumbs about what they like and want and that’s allowing us to make the shopping experience better,” she said.
One way to do that is to make sure the items they want to buy are in stock. At Sam’s, that means staying on top of roughly 6,000 items stacked on shelves in warehouse stores that average 136,000 sq. feet.
Sam’s is using, of all things, floor scrubbers to do just that. As they travel around the stores, keeping floors clean and free of debris, they’re also capturing, in real time, images of every item in the store. These scrubbers (there’s one in each store) are equipped with inventory intelligence towers that take more than 20 million photos of everything on the shelves every day.
A Brain Corp. autonomous floor scrubber, called an Auto-C, cleans the aisle of a Walmart’s store. Sam’s Club completed the rollout of roughly 600 specialized scrubbers with inventory scan towers last October in a partnership Brain Corp.
Source: Walmart
“What this means is that I can distinguish Kellogg’s Froot Loops from Kellogg’s Frosted Flakes and the depth in which they’re stocked on the shelves,” explained Bhardwaj. “And if you think about a shelf, it doesn’t always have items in the front. They could be staked in the back and then there are shadows as well.”
This is where AI and machine learning comes in. She said the company has trained its algorithms to discern the different brands and their inventory positions, taking into account how much light there is or how deep the shelf is, with more than 95% accuracy. When a product gets to a pre-determined level, the stock room is automatically alerted so that the item is always available.
If there’s nothing in the stock room, but a delivery scheduled that day contains Kellogg’s products, the algorithm will tell an associate to take that Kellogg palette directly to the sales floor, rather than to the stock room. Bhardwaj said since deploying this AI last year, employee productivity has increased 15%.
“This is how we close the loop,” she said. “We never want to be out of stock on any item.”
AI is also powering the Walmart shopping app. For example, if a customer orders Pampers on the app, it can now recognize when this customer last ordered the product and whether the size is still appropriate.
Even though Walmart has been on its AI journey for years, the goal has remained the same, Bhardwaj said: Find better ways to figure out what the customer wants to buy and how best to get it to them. “AI is a way to make those decisions super easy for us,” she said. “That’s how I look at it.”
The goal is to create as little friction between what customers want and what they ultimately buy. “I hate shopping for things like milk and toilet paper,” Bhardwaj said. “We want to make the shopping experience for everyday items a no-brainer for our customers.”
OpenAI CEO Sam Altman speaks during the Federal Reserve’s Integrated Review of the Capital Framework for Large Banks Conference in Washington, D.C., U.S., July 22, 2025.
Ken Cedeno | Reuters
OpenAI is detailing its plans to address ChatGPT’s shortcomings when handling “sensitive situations”
following a lawsuit from a family who blamed the chatbot for their teenage son’s death by suicide.
“We will keep improving, guided by experts and grounded in responsibility to the people who use our tools — and we hope others will join us in helping make sure this technology protects people at their most vulnerable,” OpenAI wrote on Tuesday, in a blog post titled, “Helping people when they need it most.”
Earlier on Tuesday, the parents of Adam Raine filed a product liability and wrongful death suit against OpenAI after their son died by suicide at age 16, NBC News reported. In the lawsuit, the family said that “ChatGPT actively helped Adam explore suicide methods.”
The company did not mention the Raine family or lawsuit in its blog post.
OpenAI said that although ChatGPT is trained to direct people to seek help when expressing suicidal intent, the chatbot tends to offer answers that go against the company’s safeguards after many messages over an extended period of time.
The company said it’s also working on an update to its GPT-5 model released earlier this month that will cause the chatbot to deescalate conversations, and that it’s exploring how to “connect people to certified therapists before they are in an acute crisis,” including possibly building a network of licensed professionals that users could reach directly through ChatGPT.
Additionally, OpenAI said it’s looking into how to connect users with “those closest to them,” like friends and family members.
When it comes to teens, OpenAI said it will soon introduce controls that will give parents options to gain more insight into how their children use ChatGPT.
Jay Edelson, lead counsel for the Raine family, told CNBC on Tuesday that nobody from OpenAI has reached out to the family directly to offer condolences or discuss any effort to improve the safety of the company’s products.
“If you’re going to use the most powerful consumer tech on the planet — you have to trust that the founders have a moral compass,” Edelson said. “That’s the question for OpenAI right now, how can anyone trust them?”
Raine’s story isn’t isolated.
Writer Laura Reiley earlier this month published an essay in The New York Times detailing how her 29-year-old daughter died by suicide after discussing the idea extensively with ChatGPT. And in a case in Florida, 14-year-old Sewell Setzer III died by suicide last year after discussing it with an AI chatbot on the app Character.AI.
As AI services grow in popularity, a host of concerns are arising around their use for therapy, companionship and other emotional needs.
But regulating the industry may also prove challenging.
On Monday, a coalition of AI companies, venture capitalists and executives, including OpenAI President and co-founder Greg Brockman announced Leading the Future, a political operation that “will oppose policies that stifle innovation” when it comes to AI.
If you are having suicidal thoughts or are in distress, contact the Suicide & Crisis Lifeline at 988 for support and assistance from a trained counselor.
WATCH: OpenAI says Musk’s filing is ‘consistent with his ongoing pattern of harassment
Okta CEO Todd McKinnon appears on CNBC in September 2018.
Anjali Sundaram | CNBC
Okta shares rose 4% in extended trading on Tuesday after the identity software maker reported fiscal results that exceeded Wall Street projections.
Here’s how the company did in comparison with LSEG consensus:
- Earnings per share: 91 cents adjusted vs. 84 cents expected
- Revenue: $728 million vs. $711.8 million expected
Okta’s revenue grew about 13% year over year in the fiscal second quarter, which ended on July 31, according to a statement. Net income of $67 million, or 37 cents per share, was up from $29 million, or 15 cents per share, in the same quarter last year.
In May, Okta adjusted its guidance to reflect macroeconomic uncertainty. But business has been going well, said Todd McKinnon, Okta’s co-founder and CEO, in an interview with CNBC on Tuesday.
“It was much better than we thought,” McKinnon said. “Yeah, the results speak for themselves.”
U.S. government customers are being more careful about signing up for deals after President Donald Trump launched the Department of Government Efficiency in January.
“But even under that additional review, we did really well,” McKinnon said.
Net retention rate, a metric to show growth with existing customers, came to 106% in the quarter, unchanged from three months ago.
Companies will need to buy software to manage the identities of artificial intelligence agents working in their environments, which should lead to expansions with customers, McKinnon said. Selling suites of several kinds of Okta software should also boost revenue growth, he said.
Management called for 74 cents to 75 cents in adjusted earnings per share and $728 million to $730 million in revenue for the fiscal third quarter. Analysts surveyed by LSEG had expected earnings of 75 cents per share, with $722.9 million in revenue. Okta expects $2.260 billion to $2.265 billion in current remaining performance obligation, a measurement of subscription backlog to be recognized in the next 12 months, just above StreetAccount’s $2.26 billion consensus.
The company bumped up its fiscal 2026 forecast. It sees $3.33 to $3.38 in full-year adjusted earnings per share, with $2.875 billion to $2.885 billion in revenue. The LSEG consensus showed $3.28 in adjusted earnings per share on $2.86 billion in revenue. Okta’s full fiscal year guidance from May included $3.23 to $3.28 per share and $2.850 billion to $2.860 in revenue.
Late last month, Palo Alto Networks, a cybersecurity company that announced an expanded partnership with Okta in July, announced plans to acquire Okta rival CyberArk for about $25 billion.
“Palo Alto is going to be like, ‘You have to buy security from us, and your endpoint from us and your SIEM [security information and event management] from us and your network from us,’ ” McKinnon said. “We just think that’s wrong, because customers need choice. It’s very unlikely they’re going to get every piece of technology or every piece of security from one vendor.”
A Palo Alto spokesperson did not immediately respond to a request for comment.
Earlier on Tuesday, Okta said it had agreed to acquire Israeli startup Axiom Security, which sells software for managing data access. The companies did not disclose terms of the deal.
As of Tuesday’s close, Okta shares were up 16%, while the technology-heavy Nasdaq was up 11%.
Executives will discuss the results with analysts on a conference call starting at 5 p.m. ET.
WATCH: We’re moving from prototype to production when it comes to agents, says Okta CEO Todd McKinnon
-
Sports3 years ago‘Storybook stuff’: Inside the night Bryce Harper sent the Phillies to the World Series
-
Sports1 year agoStory injured on diving stop, exits Red Sox game
-
Sports2 years agoGame 1 of WS least-watched in recorded history
-
Sports2 years agoMLB Rank 2023: Ranking baseball’s top 100 players
-
Sports4 years ago
Team Europe easily wins 4th straight Laver Cup
-
Sports2 years agoButton battles heat exhaustion in NASCAR debut
-
Environment2 years agoJapan and South Korea have a lot at stake in a free and open South China Sea
-
Environment11 months agoHere are the best electric bikes you can buy at every price level in October 2024