At Walmart, David Glick, senior vice president of enterprise business services at retail Behemoth, said, “Everyone uses AI across the enterprise.”
…It took several years to get enough tools, [to] Few hallucinations, [to have] Not only to make it easier than just a fewer bias. And I think the beginning of this year was when we started building agents.
In a Wednesday conversation with Atif Rafiq, founder of AI Infrastructure Startup Ritual at the AI Infrastructure Summit in Santa Clara, California, Glick explained that Walmart CEO Doug McMillon has been pushing the company all-in with AI for the past few years.
The most difficult part of that transformation was not technology. It makes people use it.
Glick said he was preparing to take the stage at the same meeting a year ago, and heard him talk in the previous session about how change management is the most challenging element of digital transformation.
He said, “I was standing behind me saying, 'No, we're engineering, engineering all the work.' That's the hardest part to actually write code. “And I was actually writing code, like I was thinking about it all day.
The problem with big companies is that everyone wants to include, but “we're driving people's cheese,” he said.
To bring everyone to the ride, Glick's goal was to let everyone use AI, not just software engineers.
“So the exciting part is not just the engineers building things,” he said. “Engineers build full enterprise-scale agents, etc. But everyone at Walmart is thinking about how to use AI at work.”
Glick has allowed only 10 or 20% of Walmart's associates to “vibe coding” or build software agents in their respective roles, as employees are called. However, he suggested that AI tools are changing what they want in technical staff.
Openai eats work and then offers to help you find something new at Walmart
read more
“As I did when I was writing the code, “Can I put a semicolon in the correct place?” or “Can I close the parentheses every time I opened it?” Glick explained. “And it was like, 'Are you malloc() and free()?” “And you say, “Oh, you don't need Java, Malloc() and Free() anymore.” And now, I think it's like curiosity, persistence, resilience, tenacity and grit. ”
Glick insists that “the future remains curious.”
He said, “What we found is the people most successful in AI, whether they're product teams, design teams, engineering teams or business teams.”
His approach at Walmart was to dive in and try things out to repeat quickly.
“I was at Walmart for over two and a half years, and since I started, we've been talking about the need to be leaders of Gen AI and Gen AI and Doug. “And it took me a few years before I could get enough of the tools. [to] Few hallucinations, [to have] Not only to make it easier than just a fewer bias. And I think the beginning of this year was when we started building agents. ”
The Glick Advocate approach is sensual, mixing the roles of developers, designers and product managers. “I don't think we even have a new slang for this,” he said. “We start with things like coding and atmosphere design and atmosphere PRD (product requirements document).
Is it a “Viber”? He wondered. designer? Even Glick's own role appears to be fluid. He said he feels like Walmart's AI chief marketing officer.
Walmart's AI bet involves changes in the business process needed to deal with AI. According to Glick, Walmart had 14 separate security processes, one of which included compliance and had a backlog of weeks or months when the project was approved.
“So we called the Chief Compliance Officer and said, “If we can build an agent in a week, it won't take two months to approve an agent,” Glick recalled. “And she went in and built a new process. It's still equally safe. It leveraged AI and used technology, but now the backlog is zero days, not 60 days.
From a change management perspective, Glick said, “I want to say I'm running the project on a stopwatch instead of a calendar, as I can turn these things while I'm sitting with a business user, and that's what I can do.
Glick stated that his goal could be to allow engineers to take laptops to shops and fulfillment centers, and before he could go home he built an agent or something that would make the life of the associates at the facility better.
“There's a lot we're trying to do,” he said. “We're trying to increase our revenues. We're trying to cut costs. But if we can change people's lives, it's better than everything.” ®
