How to get ChatGPT to recommend your products

You cannot ask ChatGPT to recommend your products, and you cannot buy your way into its answers. What you can do is make your products easy for it to find, read and match to a shopper’s question, and then measure how often that works. This guide covers how to get ChatGPT to recommend your products, what “ranking in ChatGPT” really means, and how to tell where you stand.

How does ChatGPT recommend products?

ChatGPT answers a shopping question from what it can retrieve: web pages read by OpenAI’s search crawler, structured product feeds from merchants, and, for Shopify stores, Shopify Catalog. OpenAI says reviews, rich media and performance signals help ranking, but it publishes no full formula.

Three routes bring your products into ChatGPT.

Search crawling. OpenAI’s crawler documentation says OAI-SearchBot surfaces websites in ChatGPT’s search features. Sites that opt out “will not be shown in ChatGPT search answers.” GPTBot, the training crawler, is a separate setting, so you can block training and still allow search. A third agent, ChatGPT-User, fetches a page when a person asks ChatGPT to, and OpenAI says it is not used to decide whether content may appear in search.

Product feeds. OpenAI’s commerce docs describe a structured feed with identifiers, descriptions, pricing, inventory and media. Required fields make price and availability display correctly. Recommended attributes, such as reviews and rich media, “improve ranking, relevance, and user trust.” Direct feed onboarding is open to approved partners for now.

Shopify Catalog. Shopify said in March 2026 that products become discoverable in ChatGPT by default through Agentic Storefronts. For most Shopify brands this is the main route. Our guide to Shopify Agentic Storefronts covers what you control there.

Even then, OpenAI’s feed specification is plain about the limit: “Eligibility does not guarantee display.”

How do I get my business to pop up on ChatGPT?

First make your products eligible. Allow OAI-SearchBot in robots.txt, keep Shopify Catalog access on, give every product a category, a price and an image, and keep price and stock current. Eligibility gets you considered. It does not get you chosen.

Step through the five checks below. Each one names the source that sets the rule.

Five checks before ChatGPT can recommend a product

1. 1 Crawler access

   Allow OAI-SearchBot in your robots.txt. OpenAI says opted-out sites are not shown in ChatGPT search answers, and robots.txt changes take about 24 hours to apply.

2. 2 Catalog access

   On Shopify, open Sales channels, then Agentic, and confirm ChatGPT is on. Shopify says products become discoverable in ChatGPT by default through Agentic Storefronts.

3. 3 Complete products

   Shopify Catalog leaves out products with no category, a zero price or no image. OpenAI's feed lists title, description, brand, price and availability among the fields a useful discovery feed needs.

4. 4 Current data

   OpenAI's commerce docs ask for daily feed snapshots, and Shopify Catalog updates price and stock continuously. Required fields exist so price and availability display correctly, so keep them current.

5. 5 Reviews and media

   OpenAI lists reviews and rich media as recommended attributes that improve ranking. Send real review counts and ratings only. OpenAI's spec says to omit a rating when there are no reviews.

The first check is the one most often broken by accident, usually by a robots.txt rule written for another purpose. Check yours here:

AI crawler access checker

Check whether 15 AI crawlers, from GPTBot to PerplexityBot, may read your site's pages.

Your website's address

For example yoursite.com.

Check my site

A quick check confirms you are a person, not a bot.

Descriptions checked 2026-10-13. We do not save the address or the result.

Open the full tool

How to get AI to recommend your product?

Give the assistant specific reasons to match your product to the question. State what it is, who it is for, what it is not and how it differs, answer the questions shoppers ask, and back claims with facts. Assistants match on needs, and a vague page matches nothing.

Research points the same way. A May 2026 study found that when a shopper’s question adds a need or a constraint, the recommended brands change. Your page is where you state the needs and constraints your product meets.

The original GEO paper, published at KDD 2024, tested ways to rewrite content for generative engines. It reported that GEO methods can “boost visibility by up to 40% in generative engine responses,” with adding citations, quotations and statistics among the methods that worked best. Keyword stuffing was among the methods that did not. That is a relative gain on the paper’s own metric in a research setup, not a sales figure, and the authors note that results vary by domain.

Microsoft Advertising puts the merchant version simply: whether an assistant recommends you depends on “how complete, current, and contextual your product information is.”

In practice, for each product page:

• Say what it is and who it is for. “Bar soap” matches less than “unscented bar soap for a gym bag.”
• Say what it is not. Ruling out a wrong use helps an assistant rule you in for the right one.
• Answer the real questions. Size, use, comparisons, shipping and returns.
• Back it up. Specific facts, certifications you hold and sources an assistant can repeat.

How do you rank in ChatGPT?

There is no fixed rank in ChatGPT. Each answer is generated fresh, and the same question rarely returns the same list twice. What you can track is how often your product appears across many phrasings of a question, asked many times, and on which assistants.

That share is what we call AI visibility. What is AI visibility? explains the term and how it is measured.

SparkToro had 600 volunteers run 12 prompts through ChatGPT, Claude and Google’s AI results a combined 2,961 times. The study, published January 27, 2026, found that the same list of brands came back less than 1 time in 100, and the same list in the same order less than 1 time in 1,000. Its author concluded that visibility across many prompts run many times is a reasonable metric, and that a ranking position in AI is not.

Wording matters as much as repetition. In a May 2026 paper based on about 6,000 paraphrase runs and 6,000 rerun controls on OpenAI and Anthropic models, two phrasings of the same buying need shared far fewer brands than one prompt asked twice.

How much two AI answers overlap, by how the question changed

Source: Jack et al., Paraphrase brittleness in production retrieval-augmented commercial recommendation, arXiv (May 22, 2026)

Show data

How much two AI answers overlap, by how the question changed

Pair of answers	Value (%)
Same prompt, rerun (low end of 50 to 61%)	50
Cosmetic rewording	28.8
Rewording that adds a constraint	13.5

The authors conclude that the exact wording of the prompt, more than the buyer’s underlying need, decides which brands come back. For you, that means one question tells you little. “How to rank in ChatGPT” really means: across the questions your shoppers ask, in their words, how often are you in the answer?

Be wary of any tool that reports a single “AI rank” for your product. Google says of its own AI features that “No third-party tool has access to our internal ranking or AI systems,” in its optimization guide. The honest number is a share of answers, with the questions, runs and dates behind it.

How do you know if ChatGPT recommends your products?

Ask the questions your shoppers ask, in several phrasings, many times each, across ChatGPT and the other assistants they use, and count how often your products appear. Report the result with its sample size and the range across runs, because a single answer is a sample of one.

Manual checks are a fair start, as long as you treat them as a sample, not a verdict. A simple method:

1. Write the questions in your shoppers’ words. Ten is enough to begin: what they type when they need your product, not your product’s name. Your site search, reviews and support inbox are good sources.
2. Phrase each question two or three ways. The research above shows wording changes the answer more than repetition does.
3. Ask each version several times in a fresh chat. Note every brand that comes back, not only whether you do.
4. Repeat on the other assistants your shoppers use. Each assistant can draw on different sources and give a different answer.
5. Count, then compare. Your share of answers, which competitors appear instead, and what the assistant says about you when it does name you.

You will see the variation for yourself within minutes. What a manual check cannot do well is scale: enough questions, phrasings and runs to tell a real change from noise.

One of our users, the Shopify skincare brand Jericho, shows why this comes first. When we first measured it in September 2026, ChatGPT, Gemini and Perplexity recommended it in 1 of 225 answers (0.44%). One answer on its own would have said either “always” or “never.” Neither was true.

That is the work we built SourceVision for. We ask ChatGPT, Gemini and Perplexity the questions your shoppers ask, many times over, and report how often you appear, with the range across runs. Then we draft what each product page is missing, what it is, who it is for, what it is not and a short FAQ, from facts you confirm. Nothing is written without your approval, every write is read back, and every applied change is recorded with what it replaced, so support can revert it on request. See how it works.

SourceVision is in an invite-only beta.

Request an invite

Do ChatGPT ads change which products it recommends?

ChatGPT now carries ads, bought through OpenAI’s Ads Manager, but ads are a separate product from the answer. OpenAI’s product feed keeps ad eligibility independent of search eligibility. We found no OpenAI offer to pay for a place inside organic recommendations.

OpenAI’s Ads API lets advertisers create campaigns, ad groups and ads, and every new ad goes through creative review. In the product feed, the ads flag is described as “Independent of search eligibility.” So buying ads does not make a product eligible for the organic answer, and not buying ads does not remove it.

OpenAI says that “Ads do not influence the answers ChatGPT gives you,” and that ads are labeled as sponsored and kept visually separate from the answer. So ads are a budget decision, and the organic answer still depends on your crawl access, your product data and your pages.

If you want the wider picture of AI assistants that shop and buy for people, read our guides to AI shopping agents and agentic commerce.