OpenAI and Google are betting billions that AI shopping agents will transform how people buy, but a legendary retail architect says they're solving the wrong problem. Ron Johnson, the former Apple executive who designed the company's retail store network starting in 2000, is pushing back on what he calls Silicon Valley's biggest current thesis in consumer AI: that autonomous agents will take over purchasing decisions and move shopping entirely online.

What Are Agentic Commerce Platforms, and Why Do Tech Giants Believe in Them?

Google and OpenAI have made aggressive moves into what they call "agentic commerce." Google is building the Universal Commerce Protocol, a standard designed to let AI agents guide consumers from product discovery through checkout without human intervention. OpenAI is turning ChatGPT into a shopping surface where users can research, compare, and complete purchases inside the chat window itself.

Both companies are betting that a meaningful share of retail spending will shift to these agents within the next several years. The logic seems sound: if AI can handle complex tasks like writing code or analyzing documents, why not shopping? But Johnson, now 66, argues the industry is confusing surface-level automation with genuine consumer behavior change.

Why Does Johnson Think High-Value Purchases Will Resist Automation?

Johnson's critique centers on a simple observation: people don't want to delegate major purchasing decisions to machines. When asked whether someone would let an AI agent choose and purchase a $1,000 or $2,000 laptop without ever visiting a website or a store, he was direct. A laptop is too personal a purchase to automate, he argued, because buyers want to feel the weight, judge the display, and decide which size actually fits their life.

His concession is narrower than outright rejection. Johnson expects AI to narrow the shortlist, answer questions, and pre-educate consumers before they walk through a door. He described a near future where shoppers arrive at stores already knowing what they want, a workflow shift rather than a replacement of retail. That framing puts him closer to a middle-ground view than to skepticism about AI itself.

"The secret sauce for Apple has always been its people, the people in the store, and how they treat the customer," Johnson stated.

Ron Johnson, Former Apple Retail Chief

Johnson is explicit that he is not an AI skeptic in general. He called himself "a real believer in AI" and "an AI optimist," and said he believes Steve Jobs would have embraced the technology. His disagreement is specifically with the agentic-commerce framing that treats human presence in the buying loop as a bug to be optimized away.

How to Evaluate Whether Agentic Commerce Will Actually Work

• Category: High-Consideration Purchases: Flagship items like laptops, furniture, and appliances may resist automation because buyers want sensory feedback and personalized guidance before committing to a major expense.
• Category: Low-Consideration Replenishment: Groceries, household goods, and standardized electronics represent a large share of consumer spending where the sensory experience is not the primary decision factor, making them better candidates for agent automation.
• Category: Sustained Gross Merchandise Value: The real market signal is not whether Google's Universal Commerce Protocol or ChatGPT's shopping surface can technically complete a transaction, but whether either meaningfully moves actual spending on a sustained basis and in which product categories.

The counter-case from the agentic-commerce camp is that Johnson is arguing from the category he knows best. If agents can handle even a slice of the recurring basket, the addressable market is still enormous, and Google and OpenAI are building for that slice first, not for MacBook buyers. Both companies have already demonstrated the technical ability to complete transactions; the question is whether consumers will actually use these systems at scale.

What Does Johnson's Track Record Tell Us About His Retail Predictions?

Johnson's post-Apple record complicates his current pitch. He took over J.C. Penney in 2011 with a plan to rebuild the chain from the ground up and was ousted less than two years later after sales collapsed. He now says he applied a startup mentality to what should have been a turnaround, moving too fast without bringing employees and customers along. His next act, Enjoy Technology, brought product setup to customers' homes and filed for bankruptcy in 2022, selling substantially all its assets to Asurion.

Two consecutive attempts to translate the Apple retail intuition to other contexts did not work. This history suggests Johnson's current skepticism about agentic commerce should be taken seriously, but it also means his predictions about retail transformation deserve scrutiny. If he is right that flagship purchases stay physical and human, and that AI's role compresses to a pre-visit research layer, then the agentic-commerce narrative shrinks from "the next shopping paradigm" to "a better search box," and the valuations attached to that thesis compress with it. If he is wrong, one of the sharpest retail operators of the last twenty-five years missed the transition.

The stakes for OpenAI and Google are enormous. Both companies have invested heavily in commerce features and partnerships. But Johnson's argument highlights a fundamental tension in AI development: just because a technology can automate a task does not mean consumers will accept that automation, especially when the task involves significant money or personal preference. The next few years will reveal whether agentic commerce is a genuine paradigm shift or a more modest tool that improves shopping research without replacing the human element entirely.