ChatGPT Shopping: How to Connect Your Catalog to ChatGPT Search
What OpenAI's current merchant and help docs say about ChatGPT Shopping: product results are not ads, direct product feeds exist, Shopify catalogs are already integrated, and merchant ranking depends on availability, price, quality, seller type, and checkout signals.
If you want to measure whether your catalog and merchant data are actually helping the brand appear in shopping-intent answers, GEO Scout helps track mentions, cited sources, and competitive product discovery prompts across major AI systems.
If your store appears in Google but not in ChatGPT shopping flows, the bottleneck is usually not ad spend. It is catalog structure, freshness, and merchant data.
OpenAI's current Shopping with ChatGPT Search, the live merchant page, and the Agentic Commerce Protocol now make that fairly clear.
For the broader strategy around AI visibility for retailers, see GEO for e-commerce.
What ChatGPT Shopping is and what it is not
When ChatGPT detects shopping intent, it can show products with images, pricing, product details, and merchant links.
The first important point from OpenAI's docs: product results are not ads. They are selected independently by ChatGPT, and OpenAI says partnerships do not determine which products appear.
The second important point: ChatGPT can rewrite or simplify product titles and descriptions for readability, and it can generate labels or review summaries. That means merchants should not assume the UI will mirror the raw title from the store exactly.
OpenAI also notes that:
- not every relevant product will be shown
- displayed review summaries are model-generated
- reviews and ratings are not verified by OpenAI
- prices can lag behind merchant-side updates
So the goal is not "buy placement." The goal is to make your catalog easy for ChatGPT to understand and trust.
How ChatGPT chooses products and merchants
According to OpenAI's help docs, product selection starts with user intent and conversation context. ChatGPT also considers structured metadata from first-party and third-party providers, other third-party content, model reasoning before fresh search results, and OpenAI safety policies.
Once a user clicks into a product, merchant ranking becomes a separate layer.
Merchant ranking signals documented by OpenAI
| Signal | What OpenAI says |
|---|---|
| Availability | Merchants are ranked partly by whether the item is available |
| Price | More competitive pricing can help in merchant ordering |
| Quality | OpenAI explicitly lists quality as a ranking factor |
| Seller type | OpenAI checks whether the merchant is the maker or primary seller |
| Instant Checkout availability | The help docs say this can also affect merchant ranking within the same product |
One more detail matters operationally: the first price a shopper sees is not always the lowest market price. OpenAI says the initial listing often reflects the first merchant shown, while "Best price" may appear later on the product detail page and merchant list.
Three ways to connect your catalog
OpenAI's current merchant docs point to three workable paths.
| Path | Best for | What to do |
|---|---|---|
| Crawlable website | Any merchant | Keep product and category pages accessible to OAI-SearchBot and expose clear product metadata |
| Shopify Catalog | Shopify merchants | Audit catalog quality, variant structure, and data freshness; OpenAI says Shopify data is already integrated |
| Direct product feed | Large or fast-changing catalogs | Apply to provide a direct feed through OpenAI's commerce flow and use the ACP feed specification |
1. Crawlable product and category pages
OpenAI's merchant page says any website or merchant can appear in ChatGPT search if the content is discoverable. The first technical check is simple: do not block OAI-SearchBot in robots.txt.
After that, the core work is familiar e-commerce GEO:
- unique product and category URLs
- clean titles and descriptions
- visible price and availability
- strong product images
- structured text instead of specs trapped inside images
If your public pages are weak, ChatGPT will rely more heavily on third-party providers, marketplaces, and review sites.
That is why catalog connection and on-site optimization are not separate projects. You still need strong product pages and category pages.
2. Shopify Catalog
This is the easiest route for Shopify merchants.
OpenAI's March 24, 2026 product update says Shopify merchant product data is already integrated into ChatGPT through Shopify Catalog, and the merchant page says Shopify sellers do not need extra setup or a separate application.
That changes the priority order. If you are on Shopify, do not start by building a second custom integration. Start by fixing:
- product titles
- variant naming
- image quality
- price and stock synchronization
- category and attribute completeness
In other words, the problem is usually not "how do I connect Shopify to ChatGPT?" The problem is "is my Shopify catalog clean enough to be represented correctly?"
3. Direct product feeds through OpenAI commerce
OpenAI's help docs explicitly say merchants can provide a direct product feed and apply for access. The current merchant page and ACP docs make this route much more concrete than it was a year ago.
This path is most useful when:
- the catalog changes frequently
- pricing and stock move quickly
- the assortment is large
- you want tighter control over how product data reaches ChatGPT
OpenAI's merchant FAQ also says most merchants should start with product feeds, while apps are optional and better suited to larger merchants that want more control over the shopping experience.
What the direct feed actually contains
The current ACP product spec is not just a merchant URL or sitemap. It models products and variants in a structured format.
At a high level, OpenAI's feed docs include three layers:
| Feed layer | Examples from the spec |
|---|---|
| Header | feed ID, account ID, target merchant, target country |
| Product | stable ID, product URL, description, media, variant set |
| Variant | stable ID, title, URL, price, availability, category, seller metadata |
The practical implication is important: direct feeds are about structured catalog objects, not only crawlability.
If you prepare for a feed, make sure your system can consistently expose:
- stable product IDs
- stable variant IDs
- canonical URLs
- images and other media
- current price and currency
- availability status
- seller identity
- variant attributes such as size or color
That aligns with both the shopping help docs and the current ACP product schema.
Do you need a ChatGPT app?
Usually, no.
OpenAI's merchant FAQ says most merchants should start with product feeds. Apps are optional and make more sense when you want a custom shopping or purchase experience inside ChatGPT.
For most retailers, this is the right sequence:
- Make the site crawlable.
- Clean up product and category data.
- Use Shopify Catalog if it already covers you.
- Move to a direct feed if you need more control.
- Only then consider an app.
Do not over-focus on checkout
Current OpenAI docs show that this area is evolving.
The shopper-facing help docs still describe eligible Instant Checkout experiences and say Instant Checkout availability can be one factor in merchant ranking within the same product. But the current merchant page emphasizes merchant-owned checkout on your site or app and says OpenAI is moving away from a standalone checkout experience in ChatGPT.
Inference: for most merchants, discovery data quality matters more right now than special checkout plumbing.
Common mistakes
Mistake 1: treating ChatGPT Shopping like paid placement
OpenAI's docs say product results are not ads. If the catalog is weak, money alone will not solve the visibility problem.
Mistake 2: blocking crawl access
If OAI-SearchBot cannot reach your pages, your public catalog becomes harder to surface and cite.
Mistake 3: stale price or stock data
OpenAI explicitly warns that price updates can lag. If your source data is already inconsistent, the problem gets worse.
Mistake 4: poor variant structure
If one page tries to represent too many colors, sizes, or bundles without clean variant logic, ChatGPT can display the wrong product context.
Mistake 5: building an app before fixing the catalog
Most merchants do not need a custom ChatGPT app first. They need better product data first.
Practical checklist
- Verify that
OAI-SearchBotis allowed to crawl product and category pages - Audit titles, descriptions, images, prices, and stock freshness
- Clean up variant modeling so one variant does not inherit another variant's data
- Improve product pages and category pages as public fact sources
- If you use Shopify, fix Shopify Catalog quality before building anything custom
- If you need more control, apply for a direct product feed
- Test shopping-intent prompts and see whether ChatGPT shows your products, the right merchants, and the right prices
Частые вопросы
Are ChatGPT shopping results ads?
Do Shopify merchants need a separate onboarding flow for ChatGPT shopping?
Does OpenAI support direct product feeds from merchants?
What affects merchant ranking in ChatGPT shopping?
Related Articles
How to Optimize Category Pages for AI Search: Structure, Filters, and Comparison Blocks
How to prepare category pages for ChatGPT, Perplexity, Google AI, and Alice. Which blocks to include, how to work with filters, FAQ, and category-level citable claims.
How to Optimize Product Pages for AI Answers: Structure, Schema, and Citable Claims
A practical guide to optimizing product pages for ChatGPT, Perplexity, Google AI, and Alice. What blocks to include, how to present specs, reviews, FAQ, and Product schema.
GEO for E-Commerce: How to Get Your Marketplace into AI Responses
How to optimize an online store and marketplace for inclusion in neural network recommendations. Product queries, product cards, reviews, and aggregators.