Michael Bertha

Inside Walmart’s generative AI journey

Oct 19, 20238 mins
Digital TransformationGenerative AIIT Strategy

Ben Peterson, head of People Product, shares how the Fortune One retailer went from vision to viable with its gen AI My Assistant in sixty days.

Ben Peterson, head of People Product, Walmart
Credit: Ben Peterson / Walmart

“Our people make the difference” — a common catchphrase of Walmart founder Sam Walton — still guides the company’s path forward as it ventures into the future with generative AI.

The multinational retail company positions itself as a “people-led, tech-powered” one, and sitting squarely at that intersection is generative AI, the power of which most leaders believe is fully realized when the strengths of people and technology are combined. Personifying this belief is Walmart’s new gen AI–powered My Assistant, which it launched in August to improve the productivity of its 50,000 US-based corporate employees. The move places Walmart among a handful of companies (aside from tech giants) that have leveraged generative AI at scale. And they did it in just 60 days.

A driving force behind the progress is Ben Peterson, head of Walmart’s People Product organization. After a decade-long stint at a global consulting firm, where he built a product management consulting practice serving retail and CPG clients, Peterson joined Walmart and dove straight into improving the employee experience for Walmart’s 2.1 million associates around the globe.

“We’re on an ambitious journey to reimagine the associate experience by building consumer-grade and intuitive digital experiences that unlock the fullest potential of our associates,” says Peterson, referring to the systems and processes that underpin employees’ digital interactions across hiring, onboarding, learning, performance management, career development, compensation, benefits, payroll, and more.

“Technology’s always been core to how we serve customers and support associates; and we knew early that we wanted to be a leader in generative AI,” he says.

Here, Peterson shares the genesis of Walmart’s latest innovation, as well as the secret to cutting through the hype to make good on the promise of generative AI at scale — something few organizations have accomplished to date.

Anchor on a compelling vision

“AI will not replace radiologists, but radiologists who use AI will replace those who don’t.” That’s an analogy on the lips of many leaders in Bentonville, Ark., where Walmart is headquartered. The analogy is potent: It encapsulates the company’s ethos on generative AI and reinforces the bias for innovation that enabled Walmart, years ago, to redefine supply chain management and establish its Everyday Low Prices strategy, which has been chronicled in case studies at elite business schools across the globe.

Around the time ChatGPT crossed 100 million users, Walmart executives visited a leading research institution to discuss the long-term opportunity around generative AI. The researchers confirmed Walmart’s assumptions on day one: The technology would initially primarily serve knowledge workers by augmenting their work rather than automating it. With this, Walmart’s leaders decided to focus on their first user base — the 50,000 US-based corporate “Home Office” employees.

The day after the visit to the research institution, executives convened to align on a unified top-to-top vision and messaging for generative AI at Walmart, which Peterson explains was vital amid the hype surrounding the technology.

“Between the seemingly daily new technology innovations and bifurcated perspectives across the market, it proved critical to have a clear vision that allowed us to avoid distractions and sustain progress,” he says. A clear vision not only excites and aligns; it helps teams to answer key questions and by extension mobilize for execution: How would associates access the tool? What would be the core use cases? How will our focus evolve after the first 60 days? What will the world look like 24 months from now?

Equally important, says Peterson, is your messaging across the enterprise. Walmart subdued any fear employees had of the technology by stressing from the outset that it would augment their work. “The goal is to help our associates unlock their full potential. It’s about making their work more intuitive, with more room for creativity and the human connections that make Walmart special,” he says.

Empower a cross-functional product team

My Assistant’s product-market fit was driven by myriad forces. Chief among them was buy-in from leaders who realized the potential of generative AI early on, like Donna Morris, Walmart’s chief people officer, who Peterson credits as one of the key champions in prioritizing and investing in the technology.

Until recently, Peterson’s product organization reported into the global technology   function, a common structure in organizations of Walmart’s size. While this allowed the product team to build strong relationships with other cross-functional delivery teams, its distance from the business ultimately obstructed its work. This led to a change earlier this year: moving Peterson and his People Product organization to report directly to Morris. “Given Donna’s vision, passion for technology, and strong track record of innovation, it was a critical unlock to have our team under her leadership,” Peterson explains, “and now we’re far better positioned to create world-class digital experiences that deliver outcomes for our associates and business.”

Once it came time for execution, a fully dedicated cross-functional product team was activated with high-performing leaders from product, engineering, data science, design, and the business. The team had a direct line of communication with executives when roadblocks arose, allowing them to stay laser-focused on researching, designing, and building product features that drive engagement and adoption of generative AI. Through rapid prototyping and agile development, the team developed the minimum viable product (MVP) for My Assistant and put it into employees’ hands in just 60 days.

Underpinning their product-led approach and speed to market was executive leadership alignment and participation. “Donna and other executive leaders were deeply involved in our strategy, vision, and execution, including participating in early product demos and weekly check-ins,” says Peterson, noting that their involvement was critical to maintaining focus and removing roadblocks to execution. “I can’t overstate how important executive alignment was to truly empower our product team to move with speed and mitigate the headwinds we encountered.”

Proactively manage change and training

According to Peterson, any generative AI rollout is going to encounter a change curve not unlike what Microsoft Excel experienced in the 1980s before being accepted as corporate gospel.

“Similar to how early users of Microsoft Excel had to be trained to understand how to harness the power of a PivotTable and VLOOKUP formulas, generative AI users have to understand prompting and high-impact use cases to truly harness its power,” he says, also calling attention to change management and adoption, the importance of which is not to be underestimated with regard to AI.

To drive the adoption of My Assistant, Walmart uses formal and informal tactics, conducting demos of pragmatic use cases at town halls to build awareness and then delivering hands-on training to people managers, arming them with the experience, knowledge, and confidence they need to advocate for use cases within their teams.

“We’re encouraging our leaders to talk to their teams about using My Assistant and are highlighting ways to show how it can power creativity and productivity,” explains Peterson. “If an associate needs to write a whitepaper, they shouldn’t be ashamed to use My Assistant as a starting point. Managers should celebrate its use, knowing it can lead to a more creative and productive first draft. The output won’t be perfect and will require human intervention, but My Assistant can dramatically help remove writer’s block to get started.” He says “seeing is believing” and that hands-on training will empower your people and help them see generative AI as a productivity enhancer — an ally, rather than a threat.

Looking forward

Today, My Assistant serves up a consumer-grade experience that enables Walmart corporate employees in the US to securely synthesize, summarize, and augment proprietary data from Walmart’s ecosystem. In the near term, corporate US employees will be able to use My Assistant to answer complex questions related to their benefits, as well as take advantage of other personalized use cases around career development, learning, onboarding, data analysis, and more.

Looking to the future, Walmart intends to expand My Assistant to its international corporate employees, starting with those in Canada, Mexico, and Central America. Its vision is also to eventually empower frontline associates in Walmart stores and Sam’s Clubs with gen AI solutions, to enable them to better serve customers and members.

As with any new technology, it’s hard to imagine just how much GenAI will shape our work and everyday life. What’s clear is that Peterson expects that many jobs will be augmented and, as a byproduct, many new ones created.

“But at the end of the day,” he says, “our people will continue to be what differentiates us.”

Michael Bertha

Michael Bertha is a Partner at Metis Strategy, a strategy and management consulting firm specializing in the intersection of business strategy and technology. Michael is the Head of the firm's Central Office, where he advises Fortune 500 CIOs and Digital executives on the role that technology plays in differentiating the customer experience, developing new products & services, unlocking new business models, and improving organizational operations. Prior to joining Metis Strategy, Michael spent 9 years in the IT Strategy practice at Deloitte Consulting, where he focused on working with senior leadership teams across several industries on strategic, IT-enabled business transformations. Michael has an MBA from Cornell University, and a master’s degree in the Management of IT from the University of Virginia.

More from this author