Something quietly historic happened over the 2025 holiday season. According to Salesforce's Cyber Week analysis of more than 1.5 billion shoppers, AI and AI agents influenced 20% of all global orders, driving an estimated $67 billion in sales through personalized recommendations and conversational service. One in five purchases is shaped by software acting on a shopper's behalf.
That figure marks the arrival of what the industry now calls agentic commerce: AI systems that don't just answer questions, but compare options, narrow choices, and increasingly complete purchases for the consumer. In October 2025, Walmart announced a partnership with OpenAI to let customers shop through ChatGPT, with CEO Doug McMillon declaring that the era of the search bar and the long list of results was about to change. Target, Sephora, Nordstrom, Lowe's, Best Buy, The Home Depot and Wayfair have since integrated with OpenAI's Agentic Commerce Protocol. Deloitte's 2026 retail outlook describes AI in commerce as moving from experimentation to execution, and the executives I speak to would agree. This is no longer a pilot program conversation. It is a strategy conversation.
So here is the question every retail leader should be sitting with: if an agent can research, compare and buy on a shopper's behalf, what is the store for?
I'd argue the answer is becoming clearer, and it is the opposite of what the headlines suggest.
OUR BENCHMARKS HUB 👉 Real Monthly Traffic Data From North America, EMEA, and APAC
What The Walmart Pivot Really Tells Us
The most instructive agentic commerce story so far is not a launch. It is a course correction.
By March 2026, OpenAI announced it was rethinking Instant Checkout, acknowledging that the initial version lacked the flexibility it had aspired to, and shifting its focus to product discovery while letting merchants run their own checkout experiences. Walmart, for its part, restructured the partnership to bring its own Sparky assistant into ChatGPT. As Walmart's Daniel Danker put it, shopping can start anywhere now, but customers still expect the same trusted Walmart experience wherever it begins.
Read that carefully. The world's largest retailer tested handing the transaction to an AI platform, learned that customers want consistency and trust from the brand itself, and took the experience back. Even in the most automated corner of retail, the differentiator turned out to be the relationship, not the transaction.
That is the pattern I expect to define the next five years. Agents will compress the middle of the shopping journey. The research, the comparison, the tab-hopping: much of that will be delegated. What cannot be delegated is how a brand makes a person feel, and there is no more powerful place to shape that than a physical store staffed by capable humans.
Why The Store Matters More, Not Less
Speaking on NRF's Retail Gets Real podcast from the Big Show in New York, PwC's Ali Furman and Stripe's Barbara O'Beirne described a future where physical locations serve as experiential hubs, showrooms and brand touchpoints, while some transactions happen elsewhere, sometimes invisibly through AI-driven replenishment. I read this as a promotion, not demotion, of the store.
When routine transactions migrate to agents, the purchases that remain in-store are the ones shoppers actively choose to make in person. They come for fit, for touch, for advice, for the pleasure of discovery, for a knowledgeable associate who can solve a problem no chatbot has context for. Location intelligence firm Foursquare observed heading into the 2026 back-to-school season that shoppers are visiting fewer stores with significantly higher intent, making every visit a must-win moment. Fewer visits, each carrying more weight. That is not a decline in relevance; instead, it’s a concentration of it.
There is a human truth underneath the technology story. An agent can find you the cheapest running shoe in the right size. It cannot watch you walk, notice your gait, and hand you the pair you didn't know you needed. It cannot turn a frustrated customer into a loyal one with a well-handled exchange. The more automated the rest of the journey becomes, the more those human moments stand out, and the more commercial value they carry. If agents commoditize convenience, humans differentiate on connection.
NEXT UP 👉 CRO Perspective: Nineteen Years. One Insight. Still Right.
The Store Measurement Mandate
Here is the uncomfortable part for retail leadership teams: If the store's role is shifting from transaction machine to the place where brand preference is built, then the metrics that prove store value need to shift too.
Same-store sales alone won't capture a store that inspired a purchase an agent later completed. Boards will increasingly ask harder questions: How many people came in? What did they engage with? How long did they stay? How many converted, and where did the rest fall out of the journey? What is the store's measurable contribution to demand that lands in other channels?
These have always been useful operational questions. In an agentic world, they become board-level ones, because they are the evidence that the store network is doing its new job. Retailers who can quantify traffic, engagement, and conversion with the same rigor they apply to e-commerce funnels will be the ones who can invest in stores with confidence while competitors guess.
Agentic AI is real; it is moving fast, and retailers should absolutely be building for it. But the lesson from its first year at scale, including from Walmart's very public recalibration, is that the technology amplifies the value of what it cannot replicate. Trust. Experience. Human judgment. Physical presence.
The shopper of 2026 may let an agent handle the errands. They will still walk into stores for the moments that matter, and they will reward the retailers who make those moments count. The store is not being written out of the retail story. It is being handed a bigger role, a higher bar, and a genuine opportunity for retailers ready to measure up.
MAKE YOUR STORES MORE IMPACTFUL 👉 The RetailNext Advantage