Shoppers no longer just type a product name into a search box. They ask ChatGPT, "Which is the quietest stick vacuum for a small apartment?" and a few minutes later they get a comparison guide. OpenAI calls this feature "shopping research."
In this post we explain what the feature does, where it picks products from, and how an online store can prepare for it. All the information comes from OpenAI's own announcement and help pages.
What does ChatGPT shopping research do?
According to OpenAI's announcement, shopping research starts with the user describing what they are looking for. ChatGPT first asks clarifying questions, such as the budget, who the purchase is for or which feature matters. Then it scans the web for current information such as price, availability, reviews, specs and images.
Products appear on screen while the research runs. The user can steer the search by marking them "Not interested" or "More like this." After a few minutes, a personal buying guide is ready.
According to OpenAI's help page, this guide usually includes:
- A short explanation in plain language of what to look for in this purchase
- A small number of top products, why each was chosen, their strengths and weaknesses, and links to the stores that sell them
- Comparisons that put price, size and key features side by side
- A scrollable list of other products that match the criteria
For a simple price question, a normal ChatGPT answer is enough. Shopping research is designed more for purchases that need comparison and choice.
Where does it pick products from?
The help page lists three sources:
- Product data that stores provide through the Agentic Commerce Protocol (ACP)
- Publicly available product information
- Other relevant retail sources
According to the announcement, results are organic and based on publicly available retail sites. Product pages are read directly, sources are cited, and low-quality or spam sites are avoided. The help page also says that ads in ChatGPT are separate from shopping research.
There is a similar explanation for the product cards that appear in a normal ChatGPT chat. According to OpenAI's product results page, product results are chosen independently by ChatGPT and are not ads. Structured data such as price and product description, third-party content and the user's own context play a role in the choice. Stores that sell the same product are ranked by factors such as availability, price, quality and whether the store is the product's maker or primary seller.
Is it available outside the US, for example in Turkey?
Two things need to be kept apart here. OpenAI's announcement says shopping research has started rolling out to logged-in Free, Go, Plus and Pro users on mobile and web. The announcement does not state a country limit.
The store side is different. OpenAI's product discovery page says the shopping experience where stores can share product feeds is currently live for users in the US, and that expanding to more stores and regions over time is planned.
The most reliable way to see the situation in your own category is to try it. Ask a question your buyer would ask from your own account and look at which products come up and from which sites.
What changes for online stores?
The shopper can read the guide and go straight to a store. If your product is not in the guide, how well you rank in search results may no longer matter for this shopper. So the real question is this: can ChatGPT read your product page, and when it does, does it find the right information?
1. Can it access your site?
The help page makes a clear note: some stores block automated access to their sites. In that case, shopping research skips those sources or turns to other sites that sell similar products.
This means your firewall or bot protection setting may leave you out of the guide without you knowing. OpenAI's announcement points stores that want to appear in shopping research to an allowlisting page. This page explains how to allow ChatGPT's browser traffic in services such as Cloudflare and Akamai. Also check separately whether you allow the ChatGPT search bot (OAI-SearchBot) in robots.txt.
2. Does your product page answer the question?
The guide compares information such as price, size, material and features. If this information is embedded in images on your page, or not written at all, you cannot expect the engine to find it.
Checklist:
- Are the key features written as text, in a clear list?
- Are price, stock and shipping information up to date on the page?
- Does the page say in a sentence or two who the product is for and what use it suits?
- Does your structured data match the information visible on the page?
OpenAI's product results page also says that price and shipping changes can take time to show up in ChatGPT. Outdated information can lead to a wrong comparison.
3. Can you share a product feed?
OpenAI has opened a way for stores to share product feeds directly. According to the product discovery page, catalogs of sellers on Shopify and Etsy are already integrated. Other stores can apply, and applicants are on a waitlist. OpenAI also writes that it is preparing a merchant platform during the year that will let stores connect their own data.
This door is not yet open for every country and every store. It is worth following, but today it is not a plan on its own.
4. Are you on the pages that compare you?
Shopping research reads not only store pages but also quality sources and reviews. If your name does not appear in comparison articles, review pages and guides in your category, your chance of getting into the guide may be lower. We explained how to find which pages these are in where does ChatGPT get its information.
How do you track it?
Shopping research is a long process and changes from person to person. That is why it is not easy to repeat it exactly. What you can do is ask the product questions your buyers ask regularly and track how often your brand, your products and your competitors are mentioned.
AIShortlist does this for ChatGPT, Gemini and Perplexity separately. It suggests questions from your site and does not measure until you approve them. The Sources screen shows which pages are cited as sources and which lists your brand is missing from. It does not measure shopping research itself as a separate screen; what it measures is the answers given to the same buyer questions.
Is there a guarantee?
No. OpenAI says the results are organic. No one can promise that your product will get into the guide. What you can do is remove access barriers, write product information correctly and clearly, and track the results regularly.
If you want to see where you stand in your category today, you can run a 10-question measurement with the free trial. The trial includes 1 brand and 1 measurement, and no card is required.