When Customers Stop Browsing

Online product research has long meant moving between search engines, product pages, comparison sites and brand websites. Conversational assistants can bring parts of this journey into one dialogue: a user describes a need, compares options and receives links to products and merchants. The shift is under way, but it has not removed the web or the role of online stores.

Adobe Digital Insights found that during the first quarter of 2026, traffic from AI sources to US retail websites grew 393% year on year. In March, according to the same dataset, AI-referred visits converted 42% better than non-AI visits. Adobe says its analysis is based on more than one trillion US retail visits.

These figures point to growth and different behaviour among the visitors observed. They do not describe every country, category or the whole of digital commerce; rapid percentage growth can also start from a small base. The data do not prove that users completed every decision inside an assistant. They show that the journey from AI to a merchant’s site deserves attention.

In March 2026, OpenAI announced a stronger focus on product discovery in ChatGPT. The company said the initial version of Instant Checkout did not give merchants the flexibility it wanted and that merchants would use their own payment flows, while work continued on the ACP protocol and product discovery.

This evolution is a reminder that AI-mediated commerce is not one feature. An assistant may influence which products are found and compared, while the merchant retains checkout, the customer relationship and transaction data. These are separate layers, and platforms may continue to change them.

If some product research happens in a conversation, complete and consistent product pages become more important. Materials, measurements, availability, provenance and terms should be described in ways automated systems can also read. But visibility is not control: a brand may appear in a comparison without deciding which alternatives it is placed alongside or how its identity is summarised.

Luxury faces a tension. Discovery can increase access to a brand; a comparison reduced to price and attributes can flatten differences in service, culture and meaning. The answer is not to make every product interchangeable, but to provide accurate information and authoritative sources that explain what a technical product page cannot.

When an assistant becomes a point of entry, who is responsible for how it presents a brand and the criteria it uses to select products? The answer involves platforms, merchants and the quality of information each makes available.

Adobe’s figures concern traffic to US retail sites and come from Adobe Digital Insights. OpenAI’s statements describe the platform’s own strategy. The implications for brands are an editorial interpretation, not a certain forecast of future adoption.