Retailers are placing products inside ChatGPT, Gemini, and Copilot because customers are beginning to use AI as a route to discovery and purchase. Etsy, Target, and Walmart have announced third-party partnerships, while Amazon and Walmart are also investing in their own assistants, Rufus and Sparky.
The attraction is reach. Adobe recorded a 758 percent year-on-year increase in AI-driven traffic to US ecommerce sites during November 2025 and a 670 percent rise in AI-referred visits on Cyber Monday.
The risk is that the assistant becomes the store, leaving the retailer to supply stock and fulfilment while someone else owns the decision.
Third-Party Platforms Put Products Where Shoppers Ask
Conversational commerce compresses research, comparison, and checkout into one interface. A consumer can describe a need, receive several options, refine the recommendation, and buy without navigating multiple sites.
Retailers that participate can appear at the moment of intent rather than waiting for the shopper to visit. As capabilities improve, the range of purchases handled through assistants is likely to grow from simple searches toward more complete shopping missions.
Avoiding these platforms may carry its own cost. When OpenAI introduced Instant Checkout, it suggested that enabling the function could influence merchant ranking alongside price and product quality. Catalogue participation may therefore affect both transaction access and visibility.
Retailers need product data formatted for agents rather than only human browsing. Availability, dimensions, compatibility, delivery, returns, and current price must be accurate enough for a machine to compare without inventing missing context.
The Customer Relationship Can Move Upstream
On a retailer’s own website, discovery creates behavioural data. Searches, comparisons, abandoned baskets, and navigation reveal what customers considered before buying. If those stages happen inside an external assistant, the retailer may receive an order without the story behind it.
Wharton professor Kartik Hosanagar argues that control over the agent increasingly means control over the customer relationship. Nikki Baird of Aptos raises a similar concern: even if platforms share some information, missing discovery and decision context gives retailers a poorer understanding of customers.
Deloitte reports that 81 percent of retail executives expect generative AI to erode brand loyalty by 2027. A neutral-looking assistant may prioritise price, convenience, or sponsored placement over the associations a retailer has invested in building.
The danger is not merely losing a page view. Retail media revenue, cross-selling, personalisation, and service design depend on proximity to the shopper. A retailer reduced to fulfilment has fewer opportunities to differentiate beyond stock and delivery.
Retailers Need Both Reach and a Direct Alternative
Amazon has not announced direct sales through ChatGPT and is developing Alexa+ and Rufus. Walmart combines third-party distribution with its own Sparky assistant. These mixed strategies acknowledge that no single channel is sufficient.
Retailers can use external agents for acquisition while building first-party experiences around loyalty, specialist advice, service, and post-purchase support. Contracts should define data sharing, presentation rules, ranking, fees, returns, and responsibility when an assistant gives incorrect product advice.
Store employees also need tools. A customer can already consult an AI assistant while standing in an aisle. Giving associates fast access to product and customer context may help the human service compete with a knowledgeable digital interface.
Autonomous purchasing remains the more consequential threshold. An agent representing the shopper will process product data differently from a person and may repeatedly choose based on preset rules. Retailers must learn how to serve that representative without manipulating it or losing sight of the human whose needs it is meant to express.
The transition is still early, and dramatic traffic growth begins from a small base. Retailers should test incrementally and measure customer acquisition, repeat behaviour, margin, data received, and service failures across channels.
AI commerce creates a genuine strategic tension. Refusing third-party platforms may make a retailer harder to find; accepting every platform may make the retailer easier to replace. The sustainable position is likely a portfolio: meet shoppers where they ask while retaining a direct experience worth choosing.
