GuidesGet cited by AI

How to get your products into ChatGPT shopping

ChatGPT shopping answers product questions with specific items, prices, and comparisons. To be one of them, make your product pages machine-readable, keep price and stock accurate, publish honest comparisons, and earn mentions on the sites the answers draw from. You cannot buy a placement in the organic answer, so the work is data quality and proof.

Updated November 16, 20265 steps5 min read

Check my AI visibility →

$1 today. Cancel anytime before day 7.

Getting found in ChatGPT shopping
  1. 1Fix product dataTitle, price, stock, images, and identifiers on every page.
  2. 2Add Product schemaMark up price, availability, and reviews in JSON-LD.
  3. 3Answer buyer questionsPut sizing, materials, and comparisons on the page.
  4. 4Earn outside proofReviews and roundups on sites the answers cite.
  5. 5Test the promptsRun your buyer prompts weekly and log who appears.
On this page
  1. How ChatGPT picks products
  2. Make every product page easy to read
  3. Answer the questions shoppers ask
  4. Merchant tools and feeds
  5. Mistakes that keep products out
  6. Check what ChatGPT actually says
  7. Questions

How ChatGPT picks products

Shoppers want to know whether ChatGPT can find and compare products for them, and it can: given a request like a waterproof trail shoe under a set budget, it returns a short list with reasons and prices.

Those lists come from a mix of product data feeds, pages it can read, and what reviewers and publishers say about the item. The details change, and OpenAI has been adding merchant-facing options, so check its current merchant guidance before you assume a feed is or is not required.

Whatever the mix, the inputs are the same ones you control: accurate structured data on your own pages, clear specs, and reputation on third-party sites.

Do not expect the list to match Google Shopping or an Amazon search. The answers lean toward products with plenty of independent discussion and a clear fit for the stated need, so a niche product with a precise use case can beat a famous brand on a specific prompt.

Make every product page easy to read

Start with the basics a model needs to name and compare an item: a plain product title with the brand and model, a price, availability, size or variant options, materials, dimensions, and a GTIN or SKU where one exists. Put them in text on the page, not only in an image.

Add Product schema so the same facts exist in structured form, and keep it in sync with what the page shows. A mismatch between schema price and page price is a common reason answers quote stale numbers. Product page SEO covers the on-page side.

If you run Google Shopping, the same discipline applies in Merchant Center: one clean feed that matches the site is worth more than three half-maintained ones.

Does ChatGPT mention you yet?

Checks whether ChatGPT, Gemini and Perplexity mention your brand. Tracks it every week.

Answer the questions shoppers ask

Shopping prompts are comparisons and constraints: best for wide feet, safe for a puppy, fits a 55-inch TV. Write the section of the page that answers them. A short “who it is for and who it is not for” block gives a model something specific to quote.

Then build the outside evidence. Models lean on review sites, roundup articles, Reddit threads, and YouTube, so a product that is discussed in those places gets named more often. Collect real reviews on your own pages too, and keep the ratings honest.

Comparison pages help when you have genuine differences. See comparison pages for AI for how to build them without turning into a brochure.

Product copy also needs plain facts a model can lift: dimensions in numbers, materials by name, certifications, compatible models, warranty length, and return terms. Vague lines like premium quality give a model nothing to repeat, while a spec table gives it several quotable details.

Merchant tools and feeds

OpenAI has been building ways for merchants to supply product information directly, and the programs and names have changed more than once. Rather than memorize a flow that may be out of date, read OpenAI’s current merchant documentation and check whether your platform, such as Shopify, already supports an integration.

Whichever route you take, the principle is the same one that runs Google’s shopping surfaces: one accurate source of truth. Your feed, your product pages, and your schema should say the same thing about price, availability, shipping, and returns. When they disagree, the answer will quote the one it found first, and customers will land on a page that contradicts it.

Keep identifiers consistent as well. A product that has a GTIN, a manufacturer part number, and a stable URL is easier to match against reviews and comparisons than one that exists only as a title.

A last check on the category itself. If your product sits in a crowded category, a narrow positioning, such as a specific use or customer, gives a model a reason to name you for the prompts where you are the best match, even if a bigger brand wins the generic one.

Mistakes that keep products out

The most common failure is inconsistent data: a price in the schema that differs from the page, an out-of-stock item marked available, or variants that share one title. Models and feeds both penalize contradictions, and customers notice them first.

The second is thin pages. A product page with a title, a photo, and a line of copy gives an answer engine nothing to compare. Add the spec table, the use cases, the sizing guidance, and a few real reviews with specifics.

The third is ignoring the questions after the first answer. Shoppers ask whether something ships to their country, how returns work, and whether it fits a model they own. Put those answers on the page and they become quotable lines.

Check what ChatGPT actually says

Write 15 to 30 prompts a real buyer would type for your category, run them, and record which brands and which of your products appear, with what price and which source cited. Repeat monthly at least. Answers vary between runs, so look at patterns, not one result.

Wrong prices, discontinued items, and missing variants point back to your data. Missing entirely usually means thin outside proof. If you want this tracked without a spreadsheet, Sprites can run your buying prompts through ChatGPT, Gemini and Perplexity every week and show which brands they name instead of yours.

Questions

Can ChatGPT do shopping?

Yes. It can research products, compare options against your constraints, and return a shortlist with prices and reasons. Availability of in-chat checkout depends on the merchant and region.

Can I pay to be recommended in ChatGPT shopping?

Organic recommendations are not for sale. What you control is your product data, your page content, and your reputation on the sites the answers draw from.

Do I need Product schema?

It is not a guarantee of inclusion, but it puts price, availability, and ratings in a form a machine reads without guessing. It also keeps your data consistent across Google and AI answers. See Product schema.

Which AI is best for shopping?

It depends on the category and on how fresh the data is. ChatGPT, Gemini, and Perplexity all return shortlists and sometimes disagree, so test your own prompts in each.

Keep going

See if ChatGPT recommends you

  • Checks whether ChatGPT, Gemini and Perplexity mention your brand
  • Shows which competitors they name instead
  • Tracks it every week