Q&A: Chief People Officer Donna Morris on Walmart’s embrace of generative AI

Walmart last year launched a proprietary tool powered by generative AI called My Assistant designed to assist corporate employees, or “campus associates,” with office tasks. The tool is currently only available in the US, but Walmart on Tuesday announced plans to make it available to associates in 11 other countries in their native language.

Sitting down with Retail Brew at the Consumer Electronics Show in Las Vegas, Donna Morris, EVP and chief people officer Walmart Inc., offered some insights on how the tool works and how it might be expanded in the coming years.

This interview has been edited for clarity and length.

What are some of the concrete ways that generative AI is assisting your employees?

Our desire was to build our own internal large language model that will allow our associates to do things like write a job description, prepare product specifications, use it for our merchants to assist around different items and products, and help individuals with calculations…And the way we look at it is: How can we provide the ability for our associates to move to productivity faster? And how can we optimize their time to be more human-based, where they’re getting into critical thinking skills, where they’re problem solving, where they’re working with other associates. Often many of us are burdened with a lot of tasks that get in the way from doing next-level work.

Ben Peterson, head of people product at Walmart, said recently that similar to the adoption of Excel back in the 1980s, generative AI users have to understand “prompting [the instructions that a user plugs into a large language model to get a response] and high-impact use cases to truly harness its power.” What does that entail exactly?

We’re trying to equip our associates so that they can even understand how you do effective prompting. We’ve had a bunch of boot camps where we actually talk about some of the biggest areas of work that add friction to people’s days. There’s merits in doing that, because the more you become really effective at prompting, the more it’s training the back-end model. So what Ben was trying to project is that it’s like how you have some people who are wizards today on Excel, and they’re great with macros, while the majority of people are OK with Excel, and then there are some people who will never pick it up. We’re trying to create the environment where we’ve got early adopters who become really strong at prompting, and then you’ve got the majority that learn from those that are already more advanced.

Have you considered applying this technology to non-corporate settings?

If you look at some of the productivity tools we’re releasing, those use different forms of AI and they also use different forms of [machine learning]...So the application is very specific to the tasks that they’re driving.

Has there been any push back from employees to the rollout of these AI-powered tools?

Frankly, the pandemic was a catalyst where we had to actively stand up a lot of new capabilities. We have 250,000 associates that get up every day to work on a delivery or curbside business that we never even had in 2019. And to us that demonstrates that people are willing to change provided that they’re supported in that change, and they feel like they’ve got their manager’s support and they’re being equipped to be successful.