Skip to main content
An AI shopping result is a set of product cards attached to an answer: ask a buying question and, alongside the written response, the model can show a carousel of products with prices, merchants, and ratings. Promptwatch extracts those cards from your tracked prompts’ responses, so you can measure whether your products appear, in which slot, and against whose.

​
Which models show shopping results?

Shopping extraction currently covers two surfaces:
  • ChatGPT: the shopping carousels ChatGPT attaches to buying questions.
  • Alexa: product results in Alexa’s responses.
Shopping is an enrichment on a response, not a separate response type, so coverage is per answer: the same prompt can produce a carousel one day and a plain text answer the next, which is why prompt phrasing matters so much. Setting up shopping prompts covers the phrasing that tends to produce them.

​
How the data flows in

There is no separate shopping collection. Your prompts run on their monitors’ normal schedule, and a background job sweeps new responses every 10 minutes for shopping blocks. When one is found, every card in it is stored as it appeared: product, price, merchant, rating, review count, and its position in the carousel. Those appearances fill the Shopping section in the sidebar; the Shopping section explains each chart and table there.

​
The market view and your catalog

The section splits into two tabs because “AI recommends products in my category” and “AI recommends my products” are different questions:
  • Overview: everything AI models showed for your prompts, whoever makes it. This is the market view: which products dominate the carousels, at what prices, through which merchants.
  • My products: your own catalog, matched against extracted products by the platform’s product ID (ChatGPT’s product ID or the Amazon ASIN, so upload those as your product IDs, not internal SKUs). Matching runs when you upload and again daily in the background, so a product you add today still matches when it first appears next month. See Upload and manage your products.
A response can speak warmly about your brand and still show only competitor products in its carousel, so shopping is worth tracking separately from mentions.

​
How to inspect the responses behind the cards

When you want to see the context a product appeared in, open Responses under Prompts:
  1. Filter by model (ChatGPT) and set the Enrichments filter to Shopping.
  2. Open a response to read the full answer around the carousel.
  3. Switch to the Citations tab to see the sources behind the written answer, useful context even though carousel merchants are listed on the cards themselves.

​
Availability

Shopping insights are available on Business and Enterprise plans. On other plans, the Overview tab renders demo data behind an upgrade overlay so you can see the feature; My products shows only the overlay. If you’re starting from zero: upload your catalog first, then use Generate suggested shopping prompts for your unmatched products, then watch the Overview fill as responses come in.