Walmart’s worker-led AI push runs into store-level friction

Walmart’s broad AI rollout is giving in-store workers tools meant to make daily tasks more efficient, but some employees say the systems still require frequent correction. Workers described assignments that fail to account for real store conditions, such as cleaning up spills or removing expired products before restocking shelves.

The company’s Spark delivery service has also shown how automation can create extra work. A navigation feature that tracks drivers’ real-time locations sometimes prompts them to collect items such as ice or frozen foods too early, wasting time instead of streamlining store pickups.

Walmart executives appear willing to treat those problems as part of the development process. CEO John Furner has backed an AI-for-all approach that relies on employees shaping tools from the ground up, and logistics manager Leo Garcia built an AI app that helped truckers get home faster while reducing empty trailers.

The strategy still faces practical hurdles. Companies need workers to keep offering feedback, even as AI can trigger skepticism or fear, and they must decide which suggestions to follow. Walmart’s scale, with more than 4,600 stores across the US, compounds the challenge because different locations can have sharply different needs.