Walmart conducted its first comprehensive employee AI survey in June, receiving responses from 1.7 million of its 2.1 million associates. The results offer a sharp counterpoint to national anxiety about artificial intelligence in the workplace.

Among U.S. frontline associates—the core of Walmart's labor model—84 percent reported confidence using digital tools powered by AI. Among office-based employees, 77 percent expressed excitement about AI improving their work. Associates who completed OpenAI training programs reported a 27 percent increase in confidence using AI, according to Maren Waggoner, Walmart's Chief People Officer for Associate Experience and Digital.

This diverges sharply from broader sentiment. An August 2026 Pew Research Center poll found that 52 percent of U.S. adults are more concerned than excited about AI's development.

Walmart's confidence gains rest on a specific operational choice: training and direct experience. Over 13 months, the company offered free AI certification programs in partnership with Google and OpenAI, embedding practical tools into existing workflows. The message from Waggoner was direct—providing both tools and training correlates with increased confidence.

The company's automation strategy reflects a disciplined capital allocation view: AI should augment, not eliminate, labor. In fulfillment centers, robotics handle physical tasks. At Walmart's innovation lab in Dallas, automated processes that reduced manual labor created new quality assurance roles, preventing net job losses. On store floors, customer preference for human checkout interaction reinforces the logic—checkout remains a human touchpoint.

Walmart's "people-led, tech-powered" positioning, core to its 64-year strategy, now hinges on this integration. The survey metrics suggest that workforce buy-in follows capability and clarity, not mandate. For a retailer managing margins under constant pressure, the business case is straightforward: trained associates who understand AI tools operate faster and with fewer errors. Confidence is not sentiment—it is operational efficiency.