AI Shopping

AI shopping is product discovery, research and comparison carried out through AI assistants and AI search: a shopper describes what they need in plain language and the assistant returns a short list of products with images, prices, review summaries and reasons. The purchase may then happen on the merchant's site, inside the assistant or through an agent acting for the shopper. When the AI itself completes the purchase, the topic becomes agentic commerce.

The main surfaces in October 2026 look like this:

• ChatGPT. In March 2026 OpenAI rolled out a more visual shopping experience with side-by-side comparisons to all ChatGPT users, built on product feeds merchants share through the Agentic Commerce Protocol. At the same time it wound down Instant Checkout, the in-chat purchase feature launched in September 2025, saying it lacked flexibility; purchases now complete on merchants' own sites or in merchant apps inside ChatGPT, such as Walmart's. Shopping is live for US users.
• Google. AI Mode and Gemini answer shopping questions from the Shopping Graph, which Google put at more than 60 billion listings in May 2026. Google added agentic checkout for tracked prices, and at I/O 2026 it introduced Universal Cart, a cart that works across merchants and Google services with checkout through its Universal Commerce Protocol.
• Microsoft Copilot. Copilot Checkout launched in the US on January 8, 2026, letting shoppers buy from Shopify, PayPal and Stripe merchants inside Copilot.
• Amazon. Rufus was renamed Alexa for Shopping on May 13, 2026. Amazon said in July 2026 that more than 350 million shoppers had used it over the previous 12 months, and its Buy for Me feature can purchase eligible products from other retailers' sites.
• Perplexity. It offers shopping answers with PayPal checkout, and its Comet browser agent can shop on retail sites. Amazon sued over Comet's use on Amazon.com; in August 2026 the Ninth Circuit vacated the preliminary injunction Amazon had won.

What decides which products an assistant shows differs by platform but follows a pattern. Structured product data from feeds supplies the facts: price, availability, variants, specs and policies (product feeds for AI shopping). Reviews and ratings help assistants judge quality and fit. Editorial sources, such as review sites, comparison articles and forums, shape which brands get named in the written part of the answer and often get cited. Brand recognition helps on broad queries, while detailed attributes let smaller brands win specific ones ('trail running shoes for wide feet with a rock plate').

For brands, AI shopping adds a new kind of visibility to track. A response can describe your brand warmly and still show only competitors' products in its cards, so mentions and product placements need separate measurement. Shopping features also switch on and off by query type: the ChatGPT shopping usage data shows product cards on a low single-digit share of web-search responses overall, concentrated in commercial prompts, with the rate moving in visible steps as OpenAI adjusts it. Promptwatch's shopping reports, for example, record which products appear in ChatGPT and Alexa responses, in which position and at what price.

The practical work is a mix of feed quality, Merchant Center hygiene, review generation, presence on the sources assistants cite, and tracking category prompts across platforms. Results show up in product placement rates, share of model for category prompts, and AI-referred traffic that converts on your site.