A Bet Against Themselves
There is a strange logic playing out in retail right now. The largest sellers in the world are racing to make their products available inside AI assistants they do not own, even as their own leaders say it could erode the customer relationships those products depend on. Etsy, Target, and Walmart have all moved to sell through Google's Gemini and Microsoft's Copilot, having earlier integrated with OpenAI's ChatGPT. They are wiring their catalogs into platforms that stand between them and the shopper, and they are doing it on purpose.
The numbers explain the urgency even if they do not resolve the risk. AI driven US e-commerce traffic grew 758 percent year over year in the month around Black Friday 2025, and rose 670 percent on Cyber Monday. When a channel grows at that rate, no major retailer can afford to be absent from it. The fear of being the one brand a shopper's AI assistant cannot find is, for now, stronger than the fear of what ceding the interface might cost down the line. So retailers are betting, and the bet is partly against themselves.
The Disintermediation Problem
The structural danger has a name retailers know well: disintermediation. When discovery and purchase move inside an AI assistant, the platform owns the moment that matters, and the retailer is reduced to a supplier feeding products into someone else's interface. Whoever controls the agents now has the power, warns Kartik Hosanagar, a marketing professor at the Wharton School. His warning is blunt: you are giving away the customer interactions, the brand experience, and then you become like a fulfillment company.
That is not a hypothetical. The platforms shaping these journeys also shape what gets surfaced. OpenAI's Instant Checkout can influence product ranking in search results, which creates quiet pressure on retailers to participate on the platform's terms or risk losing visibility. Aptos analyst Nikki Baird notes that retailers lose critical context when discovery and purchase happen outside their systems, the behavioral signals, the upsell opportunities, the relationship data, that they have spent decades and fortunes collecting. The interface is where the value accrues, and the interface is the thing retailers are handing over.
Why the Leaders Are Worried
The internal anxiety is measurable. In one survey, 81 percent of retail executives said they believe generative AI will weaken brand loyalty by 2027, and about half anticipate the collapse of the multi step shopping journey by the same year. These are not skeptics on the sidelines; they are the people steering the bets. When the executives building the AI commerce strategy simultaneously expect it to erode loyalty and flatten the funnel, the dissonance is the story. They are accelerating into a future they openly fear.
We think the fear is rational and the acceleration is also rational, which is what makes the situation genuinely hard. A retailer that refuses to play risks invisibility in the fastest growing discovery channel. A retailer that plays risks ceding the customer relationship to a platform. There is no obviously safe move, only a choice about which risk to run. The executives who say loyalty will weaken are not predicting their own failure; they are pricing in a structural shift they cannot opt out of and trying to position for it anyway.
The Case for Optimism
Not everyone reads the shift as loss. Kearney's Katherine Black argues that agentic commerce could deepen consumer engagement and broaden the categories people are willing to shop online, expanding the pie rather than just redistributing it. There is evidence for the upside: shoppers arriving from AI services have shown markedly higher purchase intent than those from traditional channels, which suggests the AI funnel delivers more qualified demand even if it delivers less brand control. A smaller share of a much larger, higher converting market can still be a win.
Google's framing leans into this collaborative reading. Sundar Pichai has emphasized that we only succeed together, positioning the platforms as partners that send retailers better customers rather than gatekeepers that extract rent. The optimistic case is that AI commerce is a more efficient matchmaker, connecting intent to product faster and reducing the friction that kills conversions. Whether that efficiency accrues to retailers or to the platforms that sit in the middle is the unresolved question, and the honest answer is that it depends on leverage neither side has fully established yet.
The Hedge: Build Your Own
The most telling response is from the players big enough to refuse the trade entirely. Amazon has not announced plans to sell on ChatGPT; instead it built Alexa+, its own assistant with a dedicated experience, keeping the interface and the data in house. Walmart and Amazon have both developed proprietary AI assistants, Sparky and Rufus respectively, betting that owning the agent is worth the enormous cost of building one. For retailers with the scale and engineering depth, the hedge against disintermediation is vertical integration of the assistant itself.
That option, though, is available only to a handful of giants. Most retailers cannot build a competitive assistant and must decide how to show up on platforms they will never control. For them the practical strategy is to participate while fighting to retain what they can: first party data, loyalty programs, post purchase relationships, and differentiated service that an AI summary cannot replicate. The retailers most exposed are the mid sized brands with strong products but no platform of their own, and they are the ones for whom this bet is most genuinely a gamble.
Our Read
The retail industry is making a coordinated bet on a channel that its own leaders expect to weaken the customer relationships retail has always run on. That is not irrational; it is the rational response to a discovery channel growing at triple digit rates that no major seller can ignore. But it is a bet with a real downside, and the downside is structural: the platforms that own the agent own the moment of choice, and ownership of that moment is what made retailers valuable in the first place.
Our advice to retailers is to participate without surrendering. Show up in the AI channels because the demand is real, but treat first party data, loyalty, and direct relationships as the assets to defend at all costs, because they are the only leverage that survives the shift to agentic commerce. The retailers who emerge strongest will be the ones who used the AI channel to acquire customers and then pulled them into relationships they control, rather than the ones who became, as Hosanagar warns, fulfillment companies feeding someone else's interface.
.jpg)
