How to rank in AI shopping answers: products in Google, ChatGPT and Perplexity
AI shopping answers are built mostly from structured product data, not from blog posts. Google's shopping results in AI Mode draw on its Shopping Graph, which merchants feed through Merchant Center. ChatGPT shows product results assembled from structured product metadata and, increasingly, from feeds merchants submit to OpenAI. So the first job for a store is clean, complete product data in those feeds and on the product pages, and the second is content that answers the buyer questions a feed cannot carry.
None of the engines sells placement in these organic product results, and none publishes how it orders them. What you control is whether your products are eligible, accurately described and in stock when someone asks.
Fix the feed and the product page first
- Open Merchant Center and fix everything in Diagnostics: disapproved products, price or availability mismatches, missing identifiers. Make sure free listings are enabled, since they are how products enter Google's organic shopping results.
- Fill the attributes that answer questions. GTIN, brand, MPN, color, size, material, age group, condition, shipping cost and return window. A buyer asking for "waterproof hiking boots under $150 with free returns" is filtered on exactly these fields.
- Rewrite product titles as brand, product type, then the attributes that distinguish it: "Northpeak Ridge Waterproof Hiking Boot, Men's, Wide Fit". Keep the same wording in the feed and on the page.
- Match on-page Product schema to the feed. Price, currency and availability must agree with what the page shows and what the feed says. Validate with the JSON-LD validator. Hydrogen SEO supports Product schema, set per page in its SEO Settings sheet; whichever tool prints it, view the page source and make sure only one Product block appears, since WooCommerce and some themes add their own.
- Check crawler access to product URLs. Make sure robots.txt, your CDN and security plugins let OAI-SearchBot, PerplexityBot and Bingbot reach product and category pages, not only Googlebot.
- Look at the merchant programs below and apply where you are eligible.
- Run ten shopping prompts in AI Mode, ChatGPT and Perplexity and record which products and stores appear.
Where each engine gets product data
| Engine | Main product source | What to do |
|---|---|---|
| Google AI Mode and AI Overviews | Shopping Graph, fed by Merchant Center and crawled pages | Keep the feed clean; Google lists up-to-date Merchant Center data among the basics for its AI features |
| ChatGPT | Structured metadata from third-party providers, reviews, and merchant feeds | OpenAI documents a product feed specification and an application process for merchants, tied to checkout inside ChatGPT |
| Perplexity | Its own index plus merchant data | Perplexity has run a merchant program for sharing product details; check its current terms |
| Copilot | Bing, including Microsoft Merchant Center | Submit a feed to Microsoft Merchant Center if you sell in its markets |
These programs change quickly. Check each operator's current merchant documentation before building a process around it, and keep one master product catalog so every feed carries the same facts.
What the feed cannot say
Feeds describe a product. They do not answer "will this fit a narrow foot", "is it worth the extra $40 over the cheaper model" or "how long does the battery last in cold weather". AI shopping answers pull those judgments from pages and reviews, which is where your site comes back in:
- Product pages with real detail: measurements, materials, care, what is in the box, who it is not for.
- Buying guides and comparisons that set out options by use case, with the trade-offs stated plainly.
- Size and compatibility guides in text and tables, not only as images.
- Reviews with specifics. Ask buyers about the use case, not just a star rating, and show review content on the page. Only mark up ratings that are genuinely displayed.
- Visible policies: shipping times, return windows, warranty terms. Buyers ask about them, and answers cite them.
The product page itself follows the usual rules; our WooCommerce product page guide covers it, and RankWave AI's notes on product pages cover the format for AI answers.
Tracking products in AI answers
There is no single report for AI shopping visibility. Merchant Center shows performance for your free listings, but Google has not documented a way to separate AI Mode shopping traffic from other surfaces. Links from ChatGPT carry utm_source=chatgpt.com, so filter landing pages for product URLs with that source. Visits from perplexity.ai and copilot.microsoft.com land in referral data. Pair all of this with a monthly run of the same shopping prompts, and note price and stock accuracy in each answer: a product shown at last month's price is a feed problem you can fix.
Decisions the engines keep
Engines decide which products to show and in what order, weighing factors they do not publish, and marketplaces with large catalogs and many reviews often appear ahead of single-brand stores. Agent-driven checkout is new and changing, with different protocols and eligibility rules at Google and OpenAI. Your job is to make sure that when an engine considers your product, every fact it reads about price, stock, fit and policy is correct and consistent.
Common questions
Do I need Google Ads to appear in AI Mode shopping results?
No. Organic product results come from free listings and product data Google collects, not from ad spend. Ads are a separate, labeled product.
Are ChatGPT product results paid placements?
OpenAI says its product results are chosen independently and are not ads. Merchants can supply feeds, but submitting one does not buy a position.
Does WooCommerce send products to AI engines automatically?
No. WooCommerce outputs product pages and some structured data, but feeds for Merchant Center or other programs need a feed plugin or an integration you set up.