Skip to main content
When people ask ChatGPT what to buy, the answer often includes shopping cards: product tiles with an image, a name and a link to buy. AI Shopping shows you which products appear in those cards on the prompts your workspace tracks, how often they appear, and which brand each product belongs to. You find it in the sidebar under Analysis > Shopping.
AI Shopping is in beta. KIME currently tracks shopping cards in ChatGPT only. More answer engines will be added as they launch shopping features.

​
What you see

​
Shopping over time

A line chart showing the percentage of completed ChatGPT answers that contained at least one shopping card, day by day. It tells you how “shoppable” your prompts are. A high share means ChatGPT treats your prompts as buying questions and shows products, so product visibility matters a lot for your category.

​
Top products

A table of the products that appeared most often in ChatGPT answers for the selected period and filters.

​
Product detail

Click any product in Top products to open its detail page. At the top you see three numbers:
  • Chats: how many answers included the product.
  • Appearances: how many times it appeared in total.
  • Last seen: the most recent date the product was shown.
Below is a list of every time the product appeared, with the Prompt that triggered it, the Model that answered, the Location (country) the prompt was run from, and the Date.

​
Filters

Use the bar at the top of the page to narrow the data:
  • Date range: pick custom dates, or use the 7 days, 14 days and 30 days presets.
  • Countries: see results for one market.
  • Brands: focus on products from one brand, such as your own or a single competitor.
  • Categories and Tags: limit the data to a group of prompts, for example one product line.
  • Models: limit the data to specific AI models.
The product detail page has the same filters, except Brands.

​
How to set it up

AI Shopping has no separate setup. It runs on the prompts, competitors and markets already in your workspace. To get useful shopping data, make sure that:
1

You track prompts with buying intent

Shopping cards mostly appear on questions where someone is ready to choose a product, for example “best waterproof running jackets”, “cheapest gym leggings 2026” or “which brand makes the best running shoes for women?”. Add prompts like these in Prompts.
2

Your prompts are organized with categories and tags

Group prompts by product line, price point or intent (for example “Best / Top searches” or “Affordability & price searches”). This lets you filter AI Shopping to one product area.
3

Your competitors are added

Products can only be matched to brands your workspace knows about. Add the brands you compete with in Competitors so their products show a Matched brand instead of a dash.
4

You track every market you sell in

Shopping cards and the products in them differ by country. Add your markets and languages in Markets & languages.
5

ChatGPT is one of your answer engines

Shopping cards are currently tracked in ChatGPT only, so make sure it is enabled in Answer engines.
Data starts to appear after your prompts have run. Give it a few days for the trend chart to become meaningful.

​
How to use it

1

Check how shoppable your prompts are

Start with Shopping over time. If a large share of answers include shopping cards, product visibility is a key battleground in your category.
2

See which products win

Look at Top products. Are your products in the list? Which products from competitors or untracked brands rank above them?
3

Filter to your own brand, then to competitors

Use the Brands filter to see all of your products that appear, then switch to each competitor to compare the range they get recommended.
4

Open a product to see why it shows up

Click a product to see exactly which prompts, countries and dates it appeared on. This tells you which questions a product “owns”.
5

Compare markets and time periods

Switch between countries and between the 7, 14 and 30 day views to spot where a product is gaining or losing ground.

​
Get the most value

Open the top competitor products and note the prompts they appear on. Those are the questions where your products should be recommended but are not. Prioritize your product pages, reviews and structured product data for those questions, and create tasks for them in Actions.
Products with a dash in Matched brand come from brands you are not tracking. If one keeps appearing near the top, add it as a competitor so it shows up across the rest of KIME too.
A product with many more Appearances than Chats is shown several times in the same answer, which usually means ChatGPT sees it as a strong fit. A product with few Chats but a recent Last seen date may be gaining traction.
After launching a product, changing prices or updating product pages, follow the product detail page week by week to see whether ChatGPT starts recommending it.
Filter by country to find markets where a product is recommended often and markets where it is missing. Use the gaps to plan localized content.
Where your products do not appear organically, you can buy a placement on the same prompts with Campaign Builder and Ads Manager.

​
Who it is for

  • Ecommerce and DTC brands that want their products recommended when shoppers ask ChatGPT what to buy.
  • Retailers and marketplaces that want to know which products and brands ChatGPT pushes in their categories.
  • Merchandising and pricing teams that want to see how product range and price positioning affect recommendations.
  • Agencies reporting on product visibility in AI answers for ecommerce clients.