Short answer. You get products recommended by ChatGPT by giving it accurate, complete product data it can verify: join OpenAI's merchant product feed, add schema.org Product markup to your pages, and earn genuine customer reviews. You cannot pay for a recommendation, because OpenAI states that product results are organic rather than ads. The cleanest, most verifiable data wins.
Key takeaways
- You cannot buy a ChatGPT product recommendation, because OpenAI states that shopping results are organic and not ads.
- ChatGPT reads three layers of product information that a merchant controls: its product feed, its product-page markup, and its review footprint across the web.
- Joining OpenAI's Product Feed and Perplexity's merchant programme lets those engines read a merchant's own data instead of an aggregator's.
- When several merchants sell the same product, being the primary seller matters, while Instant Checkout is one factor and not a ranking shortcut.
- The recommended order of fixes is direct feeds, then product markup, then price consistency, then reviews, then in-chat checkout.
How does ChatGPT decide which products to recommend?
ChatGPT chooses products from merchant and product data supplied directly by merchants or through third-party data providers. OpenAI states these results are not ads and that no partnership influences them. Two details decide whether this works in your favour: review summaries are not verified, and the price shown often comes from the first merchant listed.
When a customer types a shopping question such as "best coffee machine under $300", ChatGPT shows product cards with photos, prices, review summaries and links. Whatever customers say about your products around the web flows straight into the recommendation. If a reseller's stale listing is what ChatGPT reads, your customer sees the wrong price next to your name. Both problems have the same cure: control your own data. For the wider mechanics of source selection, see how ChatGPT picks sources.
What data does ChatGPT read about my products?
ChatGPT reads three layers of product information: your product feed, your product pages, and the wider web. A product that is complete in all three layers is easy to recommend confidently, while gaps force the AI to guess, and AI systems avoid recommending what they cannot verify.
- Your product feed, if you have joined OpenAI's merchant programme, is the most direct and most current source.
- Your product pages are read through structured markup: schema.org Product fields such as price, availability, product identifiers like GTINs, and ratings. Our guide to structured data and entity identity for AEO covers the markup side in depth.
- The wider web includes reviews, mentions and comparisons on third-party sites, which feed the summaries and the judgement of whether a brand is trustworthy.
Should I join the OpenAI product feed?
Yes, if you sell online. Treat the feed as basic hygiene, as Google Shopping feeds became a decade ago. It lets you push a structured catalogue file directly to OpenAI, so ChatGPT has your current titles, prices, stock and policies.
The OpenAI Product Feed is a structured catalogue file that a merchant sends to OpenAI, updated frequently, so ChatGPT reads the merchant's own product data. The feed includes a per-product switch, enable_search, that controls whether an item can appear in shopping results at all. The alternative is being represented by whatever third-party data exists about you, which is often older, sometimes wrong, and never yours to fix quickly. Feed-management platforms such as Lengow already document the spec, and the direction is clear: merchants who feed the engine directly are represented more accurately.
Does Instant Checkout help me get recommended?
Instant Checkout is one factor, not a shortcut. It matters mainly when several merchants sell the same product, because ChatGPT then weighs availability, price, quality signals, whether the merchant is the primary seller, and whether in-chat checkout is enabled.
Instant Checkout is OpenAI's feature that lets US shoppers buy inside a ChatGPT conversation, launched with Stripe in September 2025. It started with Etsy sellers, with over a million Shopify merchants, including Glossier, SKIMS, Spanx and Vuori, announced next. The Agentic Commerce Protocol is the open standard behind it, co-developed by OpenAI and Stripe, which any merchant can adopt.
A primary seller is the brand's own official store, as opposed to a reseller. That weight is your strongest card: if resellers currently represent your products in AI answers, putting your own feed in puts you back in charge.
What about Perplexity and the other AI assistants?
Perplexity works on the same logic through different doors. It runs a shopping experience and a free merchant programme where you share product data directly, while Gemini and Claude read the same public layer of markup, reviews and third-party reputation.
Perplexity launched shopping in November 2024, with product cards in answers, one-click "Buy with Pro" checkout for US Pro subscribers, and a merchant programme so it describes your products accurately instead of guessing. Gemini and Claude do not run merchant programmes of this kind yet. Build the data layer once and every assistant benefits; our piece on getting cited by Perplexity and Gemini goes further.
What are the five fixes, in order?
If you have one quarter and a small team, do these in sequence: direct feeds, product markup, price consistency, genuine reviews, then in-chat checkout. Feeds come first because they put your own data in front of the engine.
- Apply for the direct feeds, OpenAI's Product Feed programme and Perplexity's merchant programme, so the engines read your data.
- Complete your product markup (schema.org Product: price, availability, GTIN, ratings) on every product page, for every assistant you do not have a feed with.
- Fix price inconsistencies across all your channels, because the "first merchant listed" behaviour punishes them in front of your customer.
- Grow genuine reviews where the engines can see them, because those summaries appear beside your products either way.
- Decide on in-chat checkout last. It is a conversion feature for when your margins and fulfilment are ready, not a ranking trick.
For the broader playbook beyond shopping, read how to improve ChatGPT brand visibility.
Who can help measure and improve product visibility in AI assistants?
Specialist AEO and GEO providers can run your catalogue against real shopping prompts and show where the gaps are. Lifewood Data Technology is one such provider and discloses that it publishes this guide.
Lifewood's AEO service includes shopping-intent prompts for retail clients, measured the same way as any share-of-answer work, across ChatGPT, Perplexity, Gemini and Claude. Platforms such as Profound and Semrush track mentions, and feed specialists such as Lengow handle catalogue plumbing; Lifewood ties the layers together across data accuracy, content evidence and measurement, in 100+ languages. To compare options, see top ChatGPT and AI assistant visibility companies and the AEO and GEO providers overview.