To get an online store recommended by ChatGPT and Google AI Mode, make sure their search crawlers can read your product pages, mark up each product with price, availability, shipping and returns, keep a product feed in sync with the site, and write pages that answer the questions shoppers ask. There is no separate trick for AI.

Picture a mid-sized store that sells kitchen knives. A shopper asks ChatGPT for "a Japanese chef's knife under 150 euros that ships to Germany before Christmas". The assistant returns three products. The store stocks a knife that fits, but its robots.txt blocks OpenAI's search crawler, its product pages show the price only after a script loads, and the shipping countries sit in a PDF. The assistant had nothing reliable to quote, so it picked someone else.

Black Friday falls on 27 November 2026. Most of the fixes below take days, not hours, and are safer before your code freeze than during it.

How do AI assistants decide which products to recommend?

AI assistants recommend products they can find, read and trust: pages in their search index, product data from feeds, and consistent facts about price, stock and delivery. The model then matches those against the shopper's request.

The two biggest assistants say this in plain terms:

• Google AI Overviews and AI Mode. Data source: Google's normal search index. What Google says: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." The same SEO basics apply.

• Google Merchant Center. Data source: your product feed. What Google says: free listings can appear "on Google Search, Google Maps, Gemini, YouTube, the Shopping tab, Google Images, and Google Lens."

• ChatGPT search. Data source: pages crawled by OAI-SearchBot, plus product feeds that OpenAI ingests. What OpenAI says: sites opted out of OAI-SearchBot "will not be shown in ChatGPT search answers."

So the work is the same work as good product SEO, done more carefully. How to Get Your Company Cited by ChatGPT, Perplexity and Google AI Mode covers the general side of AI search; this post covers what is specific to products.

Can ChatGPT and Google actually read your product pages?

Check this first, because nothing else matters if the answer is no. Your robots.txt must allow the search crawlers, and the key product facts must be in the HTML the crawler receives.

Which crawlers to allow

OAI-SearchBot is OpenAI's crawler for ChatGPT's search features. OpenAI says robots.txt changes for it take about 24 hours to take effect. GPTBot is a different crawler, used for model training. Blocking GPTBot does not remove you from ChatGPT search; blocking OAI-SearchBot does. Many stores block both by accident with one broad rule.

For Google, AI Overviews and AI Mode only use pages that are indexed and eligible to show a snippet in Search. A nosnippet tag or a noindex on product pages removes them.

What the crawler needs to see

• Price, currency and stock status in the page source, not only after client-side scripts run.

• One canonical URL per product, so variants don't compete with each other.

• Shipping countries, delivery times and the return window as text on the page, not in an image or a PDF.

What product data should every page carry?

Every product page should carry merchant listing structured data: the product's name, image and an offer with a price and currency. Add the identifiers and policies shoppers ask about, because assistants use them to filter.

Structured data is code in the page, usually JSON-LD using schema.org vocabulary, that describes the product in a form software can read without guessing. Google's product structured data guide separates two types. Product snippets are for pages where people can't buy, such as review sites. Merchant listings are for pages where customers can buy from you.

What Google requires and recommends for merchant listings

The merchant listing documentation lists:

1. Required on Product: name, image and offers. Offers must be a single Offer, not an AggregateOffer.

2. Required on Offer: price and priceCurrency. Merchant listings need a price greater than zero.

3. Recommended on Product: a GTIN, brand, description and aggregateRating.

4. Recommended on Offer: availability, shippingDetails, hasMerchantReturnPolicy, itemCondition and url.

Google recommends putting a store-wide return and shipping policy in your Organization markup, and adding policy markup to an individual Offer only where that product differs. When both exist, the product-level policy wins.

Check your pages in the Rich Results Test and watch the Merchant listings report in Search Console. Both show missing or invalid fields before shoppers ever see the product.

Should you submit product feeds to Google and ChatGPT?

Yes, if you can keep them accurate. A feed gives the assistant your whole catalogue in one structured file, including products its crawler hasn't reached.

A product feed is a file listing every product with its attributes, such as ID, title, price, availability and link, that you send to a platform on a schedule. Google says that "providing both structured data on web pages and a Merchant Center feed maximizes your eligibility to experiences", and it may combine the two, for example using the feed price when the page doesn't show one.

OpenAI publishes its own product feed specification. It accepts JSONL or Google-compatible CSV, TSV and TXT files, and nine fields are required: item_id, title, description, url, brand, seller_name, image_url, availability and price. A flag called is_eligible_search controls whether a product can appear in search. Checkout inside ChatGPT is a separate integration, enabled only after onboarding.

Keep the page and the feed in step

An assistant that sees one price on the page and another in the feed has a reason to trust neither. Black Friday is when this breaks, because prices change several times in a week.

• Generate the feed from the same source as the site, not from a spreadsheet someone edits by hand.

• Schedule the sale price changes in your shop system, so the page, the markup and the feed switch together.

• Refresh the feed often enough to catch sell-outs; a recommended product that is out of stock costs you the shopper's trust as well as the sale.

If your feeds are stitched together across several tools, n8n vs Zapier vs Make in 2026 covers how to pick the automation layer that keeps them in sync.

How do you write product pages that answer shoppers' questions?

Write each product page so a single paragraph could answer the question a shopper would type into an assistant: what it is, who it suits, how it compares and what it costs to get it home.

Shoppers ask assistants long, specific questions. "Waterproof hiking boots for wide feet that I can return for free" contains four filters. If your page states the width options and the return terms in plain sentences, an assistant can match all four. If it only says "comfortable fit, easy returns", it can match none.

What to add to a product page

• A two-sentence summary at the top: what the product is and who it is for.

• The specifications shoppers filter by (size, material, dimensions, compatibility) as text, with units.

• A short "compared with" line for your closest alternative, including your own other products.

• Delivery times by country and the return window, written out.

• Answers to the three questions customer service hears most about this product.

• Real reviews on the page. Don't invent ratings or mark up reviews you don't show.

What should a store fix before Black Friday on 27 November?

Fix access first, then data, then content, and stop risky changes a week or two before the sale. In order:

1. Robots.txt: confirm OAI-SearchBot and Googlebot can reach product and category pages. Decide on GPTBot separately.

2. Indexing: check that best-selling products are indexed in Search Console and carry no noindex or nosnippet.

3. Rendering: view the page source and confirm price, currency and availability are in it.

4. Markup: run the top products through the Rich Results Test and fix every missing required field.

5. Policies: add Organization-level shipping and return policy markup, then product-level only for exceptions.

6. Identifiers: add GTINs where your products have them.

7. Feeds: confirm the Merchant Center feed matches the site, and decide whether to prepare a feed to OpenAI's specification.

8. Sale prices: schedule them in the shop system so the page, the markup and the feed change together.

9. Content: rewrite the summaries and FAQs of the 20 to 50 products you most want recommended.

10. Freeze: pause template and theme changes in the last week before the sale, and keep watching the Merchant listings report.

How we approach it at Vizio AI

We start with what an assistant actually receives. For the products that matter most, we fetch the page the way each crawler does, read the robots rules, extract the structured data and compare it field by field with the feed. Most gaps show up in that first comparison.

Then we fix the source, not the symptom. Prices, stock and policies come from one system and flow to the page, the markup and every feed together. Content changes follow the same pattern as the rest of our GEO work: answer first, specific facts, one product per page. After the changes ship, we ask the assistants the questions real shoppers ask and check which products come back and whether the facts they quote are right.

This is part of the work of our Growth Studio.

Sources

• Google Search Central, AI features and your website: https://developers.google.com/search/docs/appearance/ai-features

• Google Search Central, Introduction to Product structured data: https://developers.google.com/search/docs/appearance/structured-data/product

• Google Search Central, Merchant listing (Product, Offer) structured data: https://developers.google.com/search/docs/appearance/structured-data/merchant-listing

• Google Merchant Center Help, Free listings for products: https://support.google.com/merchants/answer/9199328

• OpenAI, Overview of OpenAI Crawlers: https://developers.openai.com/api/docs/bots

• OpenAI, Product feed reference: https://developers.openai.com/commerce/specs/feed

• Vizio AI, How to Get Your Company Cited by ChatGPT, Perplexity and Google AI Mode: https://www.vizio.ai/resources/blog/generative-engine-optimization

• Vizio AI, n8n vs Zapier vs Make in 2026: https://www.vizio.ai/resources/blog/n8n-vs-zapier-vs-make

• Vizio AI, Growth Studio: https://www.vizio.ai/studios/growth-studio