At Enterprise Connect AI’s first day closing keynote, Walmart’s Enterprise Business Services VP David Glick emphasized three tenets that he believes are necessary to leading a team, and a division, through technology change: focusing on operational excellence, figuring out if it’s better to buy or build a technology solution, and making sure the right people with the right skills are in the right place.
Glick’s team supports the store associates’ digital experience, finance and tech services, keeps the company safe, secure and compliant and supports the stores’ transactional systems. Since taking over the team about 18 months ago, Glick has focused on eliminating, automating and optimizing business processes. But that approach does not necessarily require AI or even generative AI.
For example, Glick optimized Walmart’s process of matching millions of invoices to purchase orders by changing one line of code. That released some of the constraints on their automated matching framework and thus “reduced the number of items our associates had to tackle manually by 30%,” Glick said.
Next, Glick and his team tackled the flood of auto-generated emails from Walmart’s HR system. Rather than use AI, his team identified 275 notifications where there was no action to be taken by the recipient. “Following this removal, we reduced the average time to take a real action in the HR system by one third,” Glick said. “So in addition to making me less annoyed, we provided real business value.”
Finally, Glick said that his team did implement Gen AI for the agents at their Benefits Help Desk. Those agents had sticky notes all over their monitors with the most frequently asked questions and answers. To solve this problem, Walmart’s 300-page benefits guide was put into a vector database. A large language model (LLM) was then trained on it. The LLM was then asked the top 100 questions received; it answered 91 of them correctly. Several days later with additional tuning, it was near 100% accuracy.
“Our original thought was to let [store] associates talk directly to it, but [then we thought] what if the AI gives the wrong answer?” Glick said. Instead, they had the contact center agents use the AI and share the answers back to the associates. “What we found was a significant reduction in average handle time and better responses to our associates.”
This is an example of Glick’s definition of operational excellence – delivering great experiences to customers by ensuring systems work the way they're supposed to all the time – in action. Although excelling at operational excellence will differ by organization, Glick recommended starting with trouble tickets. “Look at where those are coming from, what the root cause is and then classify them into the biggest and most important problems,” he said. And rather than mandate an arbitrary reduction in those tickets, “ask your teams how much they will reduce [those] tickets. That will [empower] them and help them [commit] to achieving that goal.”
Buy versus build is another critical decision faced by every organization. Implementing a vendor’s solution at an organization the size and scale of Walmart is never simple or easy. Glick said that implementing some of those solutions can take longer than it would have to build it from scratch.
“When I think about what we want to buy, I think about things that are quickly changing and have lots of content that we don't want to keep track of,” Glick said, citing the ‘canonical’ example of ever-changing tax rates and codes across all the regions in which Walmart operates. He prefers to simply call their API from his own home-built system. “That way, I get to use their content and integrate with the rest of the systems that we use – and it provides a delightful experience to our associates.”
Lastly, Glick shared an anecdote about managing a talented engineer and vice president who left for seemingly greener pastures as a CTO at a startup. That startup didn’t pan out, so the engineer expressed interest in returning to Walmart at his former level. However, there were no roles open, so Glick offered him a spot as a technical fellow. In that role, he would help teams in Glick’s organization with system architecture, doing design reviews and digging deep into the technical details. While initially ambivalent about the change, it turned out that this role exactly fit this individual’s talents – which he himself realized after being in the job for a while.
Glick shared another anecdote about having a hard conversation with an employee about their performance. “Early in my career, I was really scared of giving my people feedback, because I thought for sure they'd quit if I gave them any critical feedback,” Glick said. “Ed was super smart, but he could be very challenging. I thought he'd quit if I gave him feedback.”
Turns out, the opposite happened. After a tough, candid conversation that Glick focused around ‘what does good look like for Ed,” Ed was thrilled that Glick had taken the time to give him feedback that could help him do better.
“It's our job to build strong teams, have the right talent doing the right work, and bring authentic and transparent leadership to the table, whether it's Gen AI or the next technology we don't even know about,” Glick said. “These three things – operational excellence, knowing when to buy or build and focusing on people – will allow you to successfully lead through change.”
