What is ChatGPT Shopping?
ChatGPT Shopping is OpenAI’s product discovery and purchase feature built into ChatGPT. It allows users to search, compare, and buy products from conversational results. For Shopify merchants, organic discovery is a data-quality channel: your store must be readable, verifiable, and aligned with the buyer’s constraints.
Since early 2026, ChatGPT has been surfacing product cards with prices, images, ratings, and checkout or merchant links inside conversational answers. ChatGPT Ads now exist as a separate paid layer, but OpenAI states that shopping results are selected independently from ads. Paid media can amplify a readable product; it does not fix an incoherent catalog.
How ChatGPT Shopping works (merchant perspective)
The flow is straightforward:
- A user asks a product question - “What’s the best moisturizer for dry skin under $40?” or “I need a travel backpack with a laptop compartment.”
- ChatGPT crawls and compares available stores - it reads product pages, structured data, reviews, and policies from stores it can access.
- Product cards are displayed - each card includes the product name, price, image, star rating, and store name.
- Checkout happens on your store - the buyer is sent to your own checkout, opened in a ChatGPT in-app browser or a new tab. ACP still exists but serves product discovery: OpenAI shut down its in-chat payment step in March 2026.
The critical point: if ChatGPT cannot crawl your store, read your structured data, or verify your policies, paid placement is not a substitute for organic recommendability. Treat ads and product discovery as two different readiness problems.
The 3 layers of ChatGPT visibility
Getting your products into ChatGPT Shopping requires three distinct layers. Missing any one of them reduces or eliminates your chances.
Layer 1: Discoverability
ChatGPT uses different agents for different jobs. OAI-SearchBot supports search and shopping discovery, ChatGPT-User fetches pages for user-triggered actions, and GPTBot is a training crawler. If your robots.txt blocks the search/user agents, your store is harder to use in organic ChatGPT shopping experiences.
What to do:
- Verify that
robots.txtallowsOAI-SearchBotand does not accidentally blockChatGPT-User - Decide separately whether you want to allow training crawlers such as
GPTBot - Create a
llms.txtfile at your domain root listing your key pages - product categories, bestsellers, policy pages - Ensure your sitemap.xml is accessible and up to date
Layer 2: Readability
Once GPTBot can access your pages, ChatGPT needs to understand them. This means machine-readable structured data.
What to do:
- Implement complete
schema.org/Productmarkup on every product detail page (PDP) - Include
Offerwith price, currency, and availability in native HTML - not rendered by JavaScript only - Add AggregateRating to expose your review data (star rating, review count) in JSON-LD
- Make sure product descriptions, specs, and ingredients are in the HTML source, not loaded asynchronously
Layer 3: Transactability
ChatGPT Shopping favours stores where the purchase can be completed smoothly, and since March 2026 that completion always happens on your own checkout, not inside the chat. The handoff is the weak point to test.
What to do:
- Ensure your checkout flow works without JavaScript-dependent popups or multi-step redirects
- Publish clear, crawlable shipping and return policies as standalone HTML pages (not PDFs)
- Expose
shippingDetailsandhasMerchantReturnPolicyin your Product schema - Test the handoff itself: the ChatGPT in-app browser must reach your checkout without a popup, a login wall or a redirect chain
Step-by-step setup for Shopify
Here is the practical checklist for Shopify merchants who want to sell on ChatGPT:
-
Allow the right OpenAI agents in robots.txt - Edit your
robots.txt.liquidfile in your Shopify theme to avoid blockingOAI-SearchBotandChatGPT-User. TreatGPTBotseparately because it is a training crawler, not the same thing as ChatGPT Search. -
Complete schema.org Product on every PDP - Most Shopify themes include basic Product + Offer schema. Verify that
brand,sku,gtin,description,image, andavailabilityfields are populated. Use Google’s Rich Results Test or a JSON-LD validator. -
Expose reviews as AggregateRating in JSON-LD - If you use Judge.me, Loox, Yotpo, or Stamped, check whether the app injects AggregateRating into the page’s JSON-LD. Many review apps render stars visually but do not add structured data. Add it manually in your theme’s
product.liquidif needed. -
Create llms.txt with key pages - Add a
llms.txtfile at your store root listing: homepage, top collections, bestseller PDPs, shipping policy, return policy, FAQ, and about page. -
Ensure native HTML content - Shopify’s Liquid templates render server-side, which is good. But custom sections using JavaScript-only rendering (React widgets, lazy-loaded tabs) will be invisible to GPTBot. Critical product information - price, description, specs, reviews - must be in the initial HTML response.
-
Write clear policies as accessible pages - Shipping, returns, warranty, and privacy policies should be standalone pages (
/policies/shipping-policy,/policies/refund-policy), not embedded in PDFs or hidden behind accordions that require JavaScript to open.
ACP (Agentic Commerce Protocol) explained
The Agentic Commerce Protocol is an open standard co-developed by OpenAI and Stripe that lets AI agents - including ChatGPT - communicate with merchants to discover products, manage carts, and complete purchases. After OpenAI rolled back the in-chat Instant Checkout in March 2026, ACP now powers the Agentic Storefronts model: ChatGPT recommends products via ACP feeds and hands the buyer off to the merchant’s own checkout. The April 17, 2026 ACP spec update added cart, feed, orders, authentication, and MCP enhancements.
For Shopify merchants, Agentic Storefronts went live by default on eligible accounts in late March 2026. The preparation work is the same as before: clean checkout flows, machine-readable policies, and complete product structured data. For a deeper dive, read our full ACP guide.
Common mistakes that block ChatGPT recommendations
These are the errors we see most frequently when auditing Shopify stores for AI visibility:
| Mistake | Why it matters |
|---|---|
| GPTBot blocked in robots.txt | ChatGPT cannot crawl any page on your store |
| Prices rendered in JavaScript only | GPTBot sees an empty price field; no product card can be generated |
| Review widget without AggregateRating | Stars appear visually but ChatGPT cannot read the rating data |
| Policy pages published as PDFs | AI crawlers cannot reliably extract text from PDF files |
| Inconsistent pricing between schema and page | Price mismatch triggers a trust penalty; ChatGPT may skip the product |
| Missing product images in schema | Product cards without images get significantly lower engagement |
Measuring your ChatGPT readiness (GEO Score)
Your GEO (Generative Engine Optimization) score measures how well your store is prepared for AI-driven commerce channels like ChatGPT Shopping, Perplexity Shopping, and Google AI Mode.
A GEO audit checks all three layers - discoverability, readability, and transactability - and produces an actionable report with specific fixes for your Shopify store.
The stores that rank highest in ChatGPT Shopping are the ones that treat AI visibility as a technical discipline, not an afterthought.
For the 8 criteria ChatGPT evaluates, the actionable checklist, and a real case study (a beauty store that went from 35 to 78/100 in 10 days), see the reference guide: ChatGPT Shopping: The 8 Shopify Criteria.
Check your ChatGPT Shopping readiness: run a free GEO audit.