The Shopping report covers the products ChatGPT showed in its answers to your tracked prompts. It ranks brands by how often their products appear and names the stores that sell them. Use it when you sell products and want to know whose products ChatGPT shows.
The report is called Shopping in the Depra AI app. On this site, the same data is called product visibility.
Note: Only ChatGPT returns products. Answers from Gemini, Perplexity and Google AI Overviews are tracked for brand mentions and sources, not for product cards.
Note: OpenAI says shopping in ChatGPT is live for users in the US today and will expand to more regions (OpenAI merchants page, checked 7 Oct 2026). If your project tracks questions for India, the Shopping report may show few or no products until ChatGPT shows products there. That is not a fault. Depra AI records products whenever ChatGPT shows them in a tracked answer.
What is a shopping answer?
A shopping answer is a stored ChatGPT answer in which ChatGPT showed products. ChatGPT can show product options when a question shows that the person wants to buy. The options come with images, product details and links to sites where the shopper can learn more or buy (OpenAI Help Center, checked 7 Oct 2026). This page calls them product cards. Depra AI records the products shown in each stored answer.
An answer that names brands in its text but shows no products is not a shopping answer. Percentages in the Shopping report are over shopping answers only. Brand mentions in the answer text are measured by the visibility scores, which are explained on metrics.
What does the Shopping report show?
The report covers the tracked prompts of one project. It has a ranked list of brands, with four numbers for each brand, and three more lists. The Shopping report is on every plan.
| Part of the report | What it shows |
|---|---|
| Brands | The brands, ranked by the share of shopping answers that show their products. |
| Shopping answers | For each brand, the number of shopping answers that showed its products. |
| Share | For each brand, that number as a percentage of all shopping answers. |
| Average position | For each brand, the average position of its products. |
| Products | For each brand, the number of its products (SKUs) that appeared. SKU is short for stock keeping unit. |
| Top merchants | The stores the shown products are sold by. |
| Prompts that trigger products | The tracked prompts whose answers showed products. |
| Top products | The top products among those ChatGPT showed. |
Depra AI reads your own store's product list so that it knows which products are yours. That scan is explained on store catalogue.
What are the percentages a share of?
Every percentage in the Shopping report is a share of shopping answers, not of all answers. The base is the number of ChatGPT answers that showed products in the period you picked.
Here is an example with made-up numbers. 200 ChatGPT answers are stored for the period, and 50 of them showed products. The base is 50. A brand whose products appear in 20 of those 50 answers has a share of 40%.
Note: Read the count before the percentage. A share of 50% over 4 shopping answers is 2 answers, which is too few to call a pattern.
Which filters can I use?
The Shopping report takes the same filters as the other reports: date range, language, market, topic and tag.
- Date range: the period the stored answers come from.
- Language: questions asked in English or in Hinglish. Hinglish is Hindi and English typed in English letters.
- Market: the country the question was asked from.
- Topic: a group of prompts about one thing buyers shop for.
- Tag: a label you put on prompts, for example by product line.
Topics and tags are set on your prompts. See prompts for both, and engines and markets for the countries.
How do I see the products in one answer?
Filter the answers list to shopping answers, then open one ChatGPT answer. It shows the products ChatGPT showed and the web searches ChatGPT ran.
- 1
Filter the answers list
Set the answers list to shopping answers. It then lists only the answers in which ChatGPT showed products.
- 2
Open one answer
The full answer shows the prompt, the answer text, the tracked brands it names and the websites it cites.
- 3
Read the products and the searches
For a ChatGPT answer, Depra AI also shows the products that were shown and the web searches ChatGPT ran behind the answer.
An AI assistant can read the same data through the Depra AI MCP server, a read-only connection for AI apps. The get_shopping_summary tool returns the report, list_chats with the shopping feature lists the shopping answers, and get_chat returns one answer with its products. All three are in the MCP tools reference.
How is this different from a product prompt?
A product prompt is a buying question for one product, such as "best body balm under ₹3,000". Depra AI records whether each answer names that product, a rival's product, or neither. This works on all four engines, because it reads the answer text.
The Shopping report reads the product cards instead, so it covers ChatGPT only. A product prompt does not depend on product cards. Use product prompts for mentions on every engine, and the Shopping report for the cards.
What does the Shopping report not show?
The report covers one thing: the products ChatGPT showed for your tracked prompts. It leaves out the following.
- Other engines: only ChatGPT returns products. Gemini, Perplexity and Google AI Overviews are tracked for brand mentions and sources.
- A share of all answers: percentages are over shopping answers only.
- Marketplace rankings: top merchants are the stores ChatGPT's product cards point to. Depra AI does not track rankings inside a marketplace's own search, marketplace sales, Buy Box, ads or Amazon Rufus.
- Visits and sales: Depra AI does not measure visits or sales that come from AI answers.
- Alerts: Depra AI sends no alerts by email, WhatsApp or Slack.
- Changes to your store: Depra AI never changes a store.
AI engines can name different brands when the same question is asked twice. Treat one answer as a sample, and read the report over many answers.