Optimize your store for AI search: 8 steps

Shoppers now ask ChatGPT, Gemini, Perplexity and Google AI Mode what to buy. This guide is the whole job of optimizing an online store for AI search, in the order we would do it, for any platform: Shopify, WooCommerce, BigCommerce, Wix, Squarespace, Amazon or your own build. Each step has a short summary, a “Done when” line, a rough time and a link to the full step guide where one exists.

Steps 1 to 4 are one-time setup. Steps 5, 7 and 8 repeat, for every product and every month. You can do all of it by hand.

How do you optimize for AI search?

You optimize for AI search by making your store easy for assistants to reach, read and trust. In practice that is eight steps: open crawler access, a current llms.txt, correct product structured data, readable policies, product pages that answer questions, AI shopping programs, measurement and a monthly check.

Most of it is ordinary good practice. Google’s AI optimization guide says that, from Google Search’s point of view, optimizing for generative AI search is still SEO. What is new is the number of readers: each assistant reaches your store through its own crawler, feed or catalog, and each one needs the same facts.

For definitions, read what AI visibility is. For a list of what a ready store has in place, use the readiness checklist. This guide is the order of work.

What is GEO?

GEO stands for generative engine optimization: work aimed at getting your content and products used in answers written by AI assistants. AEO, answer engine optimization, describes the same work. Google counts both as part of SEO, so the foundations you already have still count.

Google’s guide also warns against “AEO/GEO hacks”, such as chunking content or creating unnecessary AI text files.

For a store, the difference from SEO is mostly in the result you measure. SEO aims for a ranked link. GEO aims for your product being named, and described correctly, inside the answer. Our guide to GEO vs SEO covers where the two overlap.

What are the steps to optimize a store for AI search?

Eight steps, in this order: let AI crawlers in, add or check your llms.txt, fix your product structured data, make your policies readable, answer shoppers’ questions on product pages, join AI shopping programs, measure where you stand, and keep it up to date every month.

The whole job at a glance:

The times are our rough estimates for a small catalog done by hand. A large catalog changes steps 5 and 7 most.

Step 1: Let AI crawlers in

An assistant that searches the web can only cite pages its crawler may read. OpenAI says sites that opt out of OAI-SearchBot will not be shown in ChatGPT search answers, and that each of its robots.txt settings is independent of the others. Perplexity says PerplexityBot surfaces and links sites in its search results and is not used to crawl content for AI foundation models. Google says Google-Extended does not affect inclusion in Google Search. So you can block training crawlers and still allow the search crawlers.

Check your store now. The checker reads your robots.txt and shows, crawler by crawler, whether your home, product, collection and content pages are allowed:

Done when: OAI-SearchBot, PerplexityBot, Googlebot and the other search crawlers you want are allowed on your product and collection pages. Full step: AI crawlers for ecommerce.

Step 2: Add or check your llms.txt

llms.txt is a proposed file, first published by Jeremy Howard in September 2024, that gives AI tools a short map of a site. Google’s guide says you do not need to create new machine-readable files to appear in Google Search, because Google Search itself does not use them. No assistant maker we checked says its shopping answers read one. It is still cheap to keep accurate.

Many stores already have one. Shopify manages an agents.md file by default for every store, and /llms.txt mirrors it unless your theme overrides it. Check yours, or write one for any platform:

Done when: yoursite.com/llms.txt loads, names your store and links your real collections, policies and help pages, or you have chosen to keep your platform’s default. Full step: does llms.txt work?

Step 3: Fix your product structured data

Structured data states your product facts in a form machines read without guessing. For merchant listings, Google requires a Product with name, image and offers, and an Offer with price and priceCurrency. The price must be greater than zero. Google recommends availability, brand, GTIN, shipping details and a return policy too.

Google’s AI guide says structured data is not required for generative AI search, and no special markup is needed, but calls it a good idea to keep using. Most hosted platforms output some product markup already. On Shopify, themes can print it with the structured_data filter, which outputs a Product, or a ProductGroup for products with variants. Test a product page with Google’s Rich Results Test.

Done when: a test of your main product template shows a valid Product with no errors, and the price, currency and availability in the markup match the page. Full step: Product structured data for AI search.

Step 4: Make your store policies readable to agents

An agent buying for someone needs your return window, return cost and shipping terms without guessing. Google recommends declaring a store-wide return policy once, in Organization markup, and overriding it per product only where needed. For checkout inside AI Mode and Gemini, Google requires the return cost, the return window, a link to the full policy and at least one contact method. OpenAI’s product feed has return fields too.

Done when: you have one plain returns page and one shipping page, and the same numbers appear in your markup, your Merchant Center settings and any feed you send. Full step: Return and shipping policies for AI shopping.

Step 5: Answer shoppers’ questions on your product pages

This is where assistants decide whether your product fits a question. Google’s guide asks for content that provides value beyond common knowledge. For a product page that means the facts only you have: what it is, who it is for, what it is not, and answers to the questions shoppers type.

Write those answers for shoppers and assistants, not for a search feature. Google says the FAQ rich result no longer appears in Google Search from May 7, 2026. Use only facts you can stand behind, because assistants repeat what your page says.

Done when: each priority product page answers who it is for, what it is not and its four to six most common questions. Repeat for every product. Full step: Answer shoppers’ questions on product pages.

Step 6: Get into the AI shopping programs

Crawling is one route in. Programs that take your catalog directly are the other.

• ChatGPT. OpenAI says Shopify and Etsy catalogs are already integrated, and other merchants can apply to share product data. See ChatGPT shopping.
• Google. Google says Merchant Center feeds can help your products appear in AI responses.
• Shopify. Shopify’s agentic storefronts are active by default for eligible stores and reach ChatGPT, Google AI Mode and Gemini, Microsoft Copilot and Meta. See what you control.
• Amazon. If you sell on Amazon, your shoppers meet Alexa for Shopping, which brings Rufus and Alexa+ together. For eligible products, its Buy for Me feature completes the purchase for the shopper, including on other brands’ sites. See our Alexa guide.

Done when: you know your status in each program you qualify for, and your feeds are approved with no products excluded by mistake. Full step: How to get into AI shopping programs.

Step 7: Measure where you stand

For Google, Search Console’s Generative AI performance report shows your impressions in AI Overviews and AI Mode, and Google says it rolled out to all websites as of August 31, 2026. For ChatGPT, Gemini and Perplexity there is no official report, so you ask your shoppers’ questions yourself, several times each, and count how often you are named.

Expect a low starting point. One of our users, measured in September 2026, was recommended in 1 of 225 answers (0.44%) across ChatGPT, Gemini and Perplexity.

Done when: you have a dated baseline: your question list, each assistant, the number of answers and your share of them. Repeat every month. Full step: check your AI visibility.

Step 8: Keep it up to date

AI search changes faster than most store setups. Google stopped showing its FAQ rich result in May 2026, and OpenAI says it is moving away from a standalone Instant Checkout in ChatGPT. Your own store changes too: prices, stock, policies, themes and apps. OpenAI’s feed spec asks merchants to update availability when an item sells out or comes back.

Once a month, rerun the crawler check, look at your llms.txt, test one product page’s markup, confirm your feeds updated, compare policy numbers across pages and feeds, and repeat your measurement. After any theme, app or platform change, run the first three checks the same day.

Done when: a monthly reminder is in your calendar and this month’s checks are logged. Full step: A monthly AI search checklist for stores.

You can do every step above by hand. We built SourceVision to take on the two that repeat for every product: measuring and product-page answers (steps 7 and 5). We ask ChatGPT, Gemini and Perplexity your shoppers’ questions many times over and report how often you appear, with the range across runs. Then we draft the answers each product page is missing, from facts you confirm, and nothing is written to your store without your approval. SourceVision is in an invite-only beta.

Request an invite

How to optimize content for AI search in 2026?

Write original, specific content for people, and skip the tricks. Google says its AI features depend on the same quality and spam systems as Search, that you can ignore tactics like chunking content, and that content should offer value beyond common knowledge. For stores, that means real product facts.

On a product page, commodity content is the supplier’s description that appears on twenty other stores. Non-commodity content is what only you know: who buys it and why, what it is not suited for, and what customers ask your support team. Four rules follow:

• Lead with the answer. Put the plain fact first, then the detail.
• Use the shopper’s words. Name the problem, the use and the person.
• Be specific. Sizes, materials, ingredients, compatibility and care, stated exactly.
• Keep it true. Remove claims you cannot support. An assistant will repeat them.

How can you rank your website in AI search engines?

You cannot hold a fixed rank in AI answers, because answers are written fresh and tailored to each shopper. OpenAI says ChatGPT tailors products to preferences, context and product data. What you can control is being reachable, readable and accurate, then measuring how often you are named.

A results page has positions. An AI answer has a short list that changes with the question, the shopper and the run. So you measure your share of answers across many runs, not one position.

Google sets two conditions for its own AI features. A page must be indexed and eligible to be shown in Google Search with a snippet, and the site must stay included in Search generative AI features, a Search Console control that is on by default. If a page cannot appear in Search, it cannot appear in AI Overviews or AI Mode either.

How do you optimize for ChatGPT search?

Allow OAI-SearchBot, so ChatGPT search can find and cite your pages. Then get your products into ChatGPT’s shopping data: Shopify and Etsy catalogs are already integrated, and other merchants can apply to share a feed. Keep product facts complete and current, because ChatGPT matches products to each shopper.

OpenAI says you do not need a product feed if ChatGPT already crawls your site, but a feed gives you more control, and richer product data helps your products surface with more accurate information. Shopping in ChatGPT is live for users in the U.S.

The full walk-through is in our guides to ChatGPT shopping and to getting ChatGPT to recommend your products.

How to optimize for AI Overviews?

Treat it as Google SEO with care for product data. A page must be indexed and eligible for a snippet. Google needs no special files or markup for its AI features, says Merchant Center feeds can help products appear in AI responses, and reports AI Overviews impressions in Search Console.

For a store, that comes down to three checks:

1. Indexing. Your product and collection pages are indexed, with Googlebot allowed, and Search Console’s Search generative AI control still includes your site.
2. Merchant Center. Your products are approved and current in Merchant Center.
3. Measurement. Search Console’s Generative AI report shows which pages earn impressions.

Work through the eight steps in order, then come back to steps 5, 7 and 8 every month.