McMillon Handed Off Before He Could Finish What He Started. That’s the Point.

McMillon told CNBC in December that he had decided to step down because of what he could see coming with AI, not because he had finished building toward it. “With what’s happening with AI, I could start this next big set of transformations with AI, but I couldn’t finish,” he said. That was not a confession. It was a design decision, and suppliers should read it as one.

McMillon, who led Walmart since 2014, said he began thinking about succession roughly a year before his announcement when the shape of what he called “agentic commerce” came into focus. John Furner succeeded him as President and CEO on February 1, per Walmart’s official announcement. Furner had run Walmart U.S. since 2019 and is the first CEO in the company’s modern era to have overseen the full arc of its digital buildout before taking the top job.

The Coca-Cola succession CNBC reported Thursday follows the same logic. James Quincey told CNBC that his company had made progress in what he called “a pre-AI, pre-gen-AI mode” but that the next chapter required someone with the energy to pursue an entirely new transformation. Two unrelated CEO transitions, both explained publicly in terms of AI readiness, in the same news cycle, are not a coincidence. They frame a question every major supplier account team should be sitting with: if the people who built the current operating model handed it off because they could not finish the next one, what does that mean for the commercial relationships built on top of it?

The Transition McMillon Described Is Already Operational

What makes McMillon’s statement more than a valedictory observation is that the transition he described has a specific shape inside Walmart’s structure, and Furner moved quickly to make it visible.

When Furner announced a C-suite restructuring in January, he said Walmart was “centralizing our platforms to accelerate shared capabilities,” per a company statement reported by Retail Dive. The executive moves were directional: Seth Dallaire, who built Walmart’s advertising, subscription, and data businesses inside Walmart U.S., was elevated to chief growth officer for the entire company. David Guggina, an e-commerce and supply chain executive, took over Walmart U.S. Walmart Connect, Walmart+, and Walmart Data Ventures now report into a single enterprise-level function. That is not an organizational detail. It is a structural bet that the next competitive advantage comes from the intersection of those platforms, not from any one of them in isolation.

On January 11, Walmart and Google announced a partnership to embed Walmart’s shopping experience directly into Google’s Gemini app using what the companies called the Universal Commerce Protocol, per Walmart’s corporate announcement. Under that framework, Gemini surfaces Walmart and Sam’s Club products in response to conversational queries, not keyword searches, not sponsored placements in a traditional sense, but AI-generated recommendations tied to a shopper’s stated context. Planning a camping trip, managing a household, preparing for a season: the agent decides what to show. Furner, speaking at the announcement, described the shift from traditional search to agent-led commerce as “the next great evolution in retail.”

That description is now being tested against commercial reality.

The Early Agentic Commerce Data Exposed a Gap the New Model Has to Close

Walmart’s own experience with OpenAI’s Instant Checkout provides the most instructive near-term data point. Walmart made approximately 200,000 products available for purchase inside ChatGPT without users leaving the interface, Walmart confirmed to CNBC. The results did not match the architecture’s promise. CNBC reported that internal data showed conversion rates inside ChatGPT ran roughly three times lower than transactions completed on Walmart’s own website. Walmart EVP of AI acceleration Daniel Danker called the experience “unsatisfying” in remarks to Wired.

Walmart is not retreating from the agentic model. It is rebuilding it on its own terms. The retailer plans to deploy Sparky, its in-house AI shopping assistant, inside ChatGPT and Gemini with checkout remaining entirely on Walmart’s own infrastructure, per CNBC’s March 20 reporting. Under that structure, users interact with Sparky inside the third-party AI environment but log directly into their Walmart accounts, and the cart reflects ongoing shopping behavior rather than a single session. Speaking at Morgan Stanley’s Technology Media and Telecom Conference in March, Danker said purchases associated with Sparky use in ChatGPT showed signs of additive volume rather than simple channel substitution, Digital Commerce 360 reported.

For suppliers, the conversion gap and the subsequent architectural pivot carry an implication that deserves attention independent of the CEO transition story. The products Walmart made available inside ChatGPT were a subset of its catalog. The AI surfacing those products operated on different inputs than the algorithm that ranks results on walmart.com. That gap, between what drives discovery in traditional search and what drives surfacing in an agent response, is where the next competitive sorting will happen.

Content Built for Search Rank Is Not Content Built for an Agent

The implications differ between 1P suppliers and 3P Marketplace sellers, and treating them as the same problem will cost teams the wrong kind of preparation time.

For 1P suppliers, the agentic surfacing question is fundamentally a content and data quality question. AI agents drawing on Walmart’s catalog to generate recommendations will weight structured product attributes, use-case relevance, and richness of item-level data more heavily than keyword density or bid levels. Item content built in Supplier One to satisfy a search algorithm is not necessarily built to satisfy an agent working from a conversational query about a shopper’s specific context. Suppliers whose content teams are still operating against traditional SEO logic for walmart.com should be asking what their item data looks like when it becomes the input to an AI recommendation rather than a search rank signal. Those are different optimization problems, and right now most supplier content strategies are only solving one of them.

For 3P Marketplace sellers, the picture is different and in some respects more urgent. Walmart’s alignment with open commerce protocols and third-party AI platforms may accelerate the emergence of shopping agents that treat retailers more like interchangeable endpoints, where price, availability, and delivery speed matter more than brand loyalty, as PYMNTS observed in January analysis of Walmart’s Google partnership. Sellers who compete on price and in-stock rate are structurally suited for that environment. Those who rely on brand-level differentiation within walmart.com search may find that new discovery surfaces produce less predictable results. WFS participation and fulfillment reliability will carry increasing weight in AI-mediated recommendations as those systems develop, and sellers who have deferred that investment are carrying more exposure than they may have priced in.

The Board Already Told You What Furner Is For

When Walmart’s board announced Furner’s appointment, Chairman Greg Penner described the transition as guiding Walmart into “a new retail era fueled by innovation and AI,” per Walmart’s official announcement. That language appears in succession announcements routinely. What gives it weight here is the structural evidence behind it: the C-suite reorganization centralizing advertising and data, the Gemini partnership built on an open commerce standard, and the Sparky rebuild now underway.

Furner told the NRF 2026 audience in January that Walmart was willing to change “what we sell, how we interact with customers, with our associates,” per Retail Dive, framing the company’s purpose and values as the fixed point and everything else as open to reconfiguration.

That is a wider aperture than most suppliers have planned around. The line review calendar, the JBP structure, the Connect investment model: those exist inside a retail operating model that Furner just described as mutable. Suppliers whose planning assumptions are built on the stability of that model should be asking which of those assumptions the succession story actually puts in motion.

Walmart U.S. e-commerce grew 27% in the fourth quarter of fiscal 2026, the eighth consecutive quarter of growth above 20% for the segment, per Walmart’s Q4 earnings release. Walmart Connect grew 41% in the U.S. in the same quarter, the company reported. The platform McMillon built is running. What Furner inherits is that infrastructure. What he is building toward is the commercial layer that runs on top of it, and the suppliers who understand that distinction have a narrowing window to act like it.

Winning With Walmart

Winning With Walmart is an independent platform for suppliers, sellers, solution providers, and industry experts.

Articles are developed by staff editors, contributing experts, and trusted partners.

Some are researched and drafted with the assistance of AI tools and are reviewed, fact-checked, and edited by Winning With Walmart editors before publication.

Editorial judgment, sourcing decisions, and final approval rest with the publication's human editors in every case.

We are committed to accuracy and fairness. If you believe this article contains an error, we welcome your feedback.