Walmart’s AI Platform Strategy: Owning the Shopping Journey Across AI Interfaces
Earlier this month I interviewed Hari Vasudev, Chief Technology Officer for Walmart U.S., at the National Retail Federation’s Big Show in New York. I wanted to press him not only on the day’s news, a new open approach to AI for commerce in partnership with Google; but also on the impacts he sees for Walmart’s massive in-store workforce.
This is an AI-assisted summary of our 30-minute conversation.
If you’d like to join me – and peers – for deeper conversations on innovation and leadership, get on this list for Fortt Knox Executive Communities, launching soon: mba.fortt.com.
Walmart’s AI strategy starts with a platform-first view of commerce rather than a bet on any single AI interface. Vasudev describes an evolution from predictive AI to generative and now agentic systems, unified by internal platforms that serve customers, associates, sellers, and developers at scale. That architecture underpins Walmart’s partnerships with both OpenAI and Google, but with a key distinction: Walmart wants the entire shopping journey, not just item-level handoffs.
The Gemini integration announced with Google embeds a full Walmart experience inside the AI interface, allowing customers to move from intent to basket-building to fulfillment with less friction. By contrast, current ChatGPT integrations focus more narrowly on single-item fulfillment. Vasudev frames this not as a platform war, but as customer-centric design: whoever reduces friction, preserves trust, and accelerates completion wins.
Crucially, Walmart is pairing openness with control. A hybrid, multi-cloud and multi-model approach gives Walmart leverage over cost, data governance, and speed of innovation. Identity and context — who the customer is and what they’re trying to do right now — become the new “currency” of commerce, replacing older constructs like cookies alone. Throughout, Vasudev emphasizes guardrails, transparency, and human-in-the-loop systems as foundational, not optional.
“It’s not just about owning the customer data. It’s equally about ensuring that we’re able to create an experience for them that is super low friction, that is essentially highly simplified, allows them to complete the journey in a very timely manner, offering them the best assortment at the best prices, with the fastest delivery speeds.”
Walmart has embedded AI across search, personalization, forecasting, and now agentic workflows
The company built internal AI platforms to serve customers, associates, sellers, and developers
Gemini enables a full Walmart experience inside AI, including basket-building
ChatGPT integrations today focus more on single-item fulfillment
Walmart prioritizes reducing friction over controlling a single interface
Its AI platform strategy mirrors its hybrid, multi-cloud architecture
Customer identity plus real-time context drive agent behavior
Human-in-the-loop design is central to trust and governance
As the conversation moves from AI platforms and abstract agents to the tangible realities of retail, the question becomes less about where intelligence lives and more about how it shows up. For Walmart, that means translating AI from digital discovery into physical stores — without losing the human connection that defines in-person retail.
AI in the Physical Store: Empowering Associates Without Replacing Them
Vasudev is clear that Walmart’s physical-store AI strategy is not about replacing people with robots, but about making associates more effective. He frames associates as customers of Walmart’s software, deserving tools that reduce friction and surface the “next best action” based on role, permissions, and real-time location. AI helps route tasks to the right associate at the right moment, minimizing wasted movement and improving responsiveness on the sales floor.
Managers, in turn, move up the stack from micromanaging tasks to higher-level planning, staffing, and inventory flow. Walmart pairs these tools with training and continuous feedback, avoiding the common enterprise failure of overwhelming workers with disconnected systems. Engineers regularly spend time in stores, using the same tools as shoppers and associates to refine design.
On automation and robotics, Vasudev draws a sharp line between purpose-built technology and humanoid replacement. Computer vision, for example, detects spills and enables frictionless exit at Sam’s Club, removing uncomfortable interactions rather than human warmth. But robots won’t clean the spills, people will. The lesson from self-checkout is balance: transactional friction can be automated, but emotionally positive moments still belong to people.
“Our philosophy as a company is we are a people-led enterprise; tech-powered, but people-led. So we strongly believe that, particularly in these early days of agentic AI, having appropriate guardrails in place, having a human-in-the-loop system is very important.”
Highlights
Associates are treated as primary users of Walmart’s AI systems
“Next best action” tools guide associates based on role and location
Managers shift toward planning, staffing, and optimization
Training and upskilling accompany every new AI deployment
Engineers regularly work inside stores to inform product design
Computer vision improves safety and enables frictionless checkout
Walmart avoids humanoid robots in customer-facing roles
Automation is used where it removes friction, not trust
If you’d like to join me – and peers – for deeper conversations on innovation and leadership, get on this list for Fortt Knox Executive Communities, launching soon: mba.fortt.com.