How to get your products into AI shopping answers (ChatGPT, Google AI Mode, Perplexity)

A growing share of product discovery starts with a question to an assistant: "best waterproof walking boots under £150", "which air fryer for a family of five", "is this coffee machine any good". The answer names products with prices and links, drawn from a handful of sources. Being in that handful requires structured, current, readable product data, and most independent shops have not put it in place.

This guide is written for a UK online shop of any size. Check the vendor documentation linked below before you start; the shopping side of AI search changes quarterly.

Step 1: Fix crawler access and rendering first

Allow OAI-SearchBot, PerplexityBot and Bingbot, check your CDN is not blocking them, and confirm that product name, price, availability and description appear in the raw HTML of a product page (View Page Source). Many themes and apps inject price and stock by JavaScript, which is invisible to most AI crawlers. Our JavaScript guide covers the fix per platform.

Step 2: Submit an accurate feed to Google Merchant Center

Create or verify your Merchant Center account, connect your shop (Shopify, WooCommerce, BigCommerce and Magento all have native connectors), and submit a product feed with complete attributes: title, description, price, sale price, availability, GTIN or MPN, brand, condition, image, shipping and returns. Fix every disapproval; disapproved products are ineligible. Google's guide on generative AI features points merchants here for a reason: it is the direct route into AI Overviews and AI Mode shopping results.

Step 3: Submit to Bing Merchant Center and OpenAI's merchant programme

Bing Merchant Center accepts the same feed format and feeds Copilot and Bing's shopping results, which ChatGPT's search draws on. OpenAI runs a merchant programme with its own product feed specification so that ChatGPT can show your products with accurate price, availability and attributes, and in some markets complete purchases. Read the current Product Feed Spec on OpenAI's developer site, prepare the feed (CSV or JSON; most shops can adapt the Merchant Center export) and register through the merchant portal at chatgpt.com/merchants. Shopify stores are already connected through Shopify's catalogue integration, so check what is flowing before building a separate feed. Requirements and availability by country change; check before assuming.

Step 4: Add Product schema that matches the feed

Every product page gets a Product block with name, image, description, brand, sku, gtin where you have it, aggregateRating and review from real reviews only, and offers with price, priceCurrency, availability, itemCondition and url. Price and availability must match the visible page and the feed. Variants (size, colour) need their own offers or a ProductGroup. Validate with the Rich Results Test.

Step 5: Write descriptions that answer buyer questions

Manufacturer copy is duplicated across every retailer and adds nothing an engine can prefer you for. Rewrite the top 50 products with the attributes buyers ask about: dimensions, weight, materials, compatibility, what is in the box, who it suits and who it does not, how it compares with the model above and below. Add a short Q&A on the page from real customer questions. Specifics are what get quoted.

Step 6: Collect reviews where engines read them

On-site reviews with Review schema, plus Google customer reviews via Merchant Center, plus Trustpilot or Feefo if your sector uses them. Marketplace reviews count too if you also sell there. Reviews with specific detail ("fits a 30-litre rucksack, arrived in two days") are matchable to questions; star ratings alone are not.

Step 7: Publish comparison and buying-guide content

"X vs Y", "best [category] for [use]", "how to choose a [category]" pages that name products (including ones you do not sell, honestly) with tables of attributes and prices. These pages match the shape of shopping questions and are heavily cited. Link each product mentioned to its page.

Step 8: Measure it

Add the AI channel group in GA4 (see our guide) and watch which product and guide pages receive AI referrals. Run a monthly prompt set of ten to twenty shopping questions across ChatGPT, Google AI Mode, Perplexity and Copilot and record which products and retailers are named. Check Merchant Center for feed health weekly.

How to know it worked

Your products appear with correct prices in AI shopping answers for your core queries. Merchant Center shows zero disapprovals. GA4 shows AI Search sessions landing on product pages and converting. Assistants quote your own page rather than a marketplace listing for your products.

Common mistakes

• Prices in the feed that differ from the page or the schema. Engines and Google both penalise mismatches.
• Skipping Bing and OpenAI because Google is "enough".
• Manufacturer descriptions on every product.
• Client-side rendered prices that only Google can see.