Generative engine optimization (GEO) is the work that gets your online store named and linked in AI answers from ChatGPT, Gemini, Perplexity, Copilot and Google AI Overviews. For a store it means four jobs: crawlers can read your pages, product data is complete, category and product pages answer buyer questions, and other sites mention you. You measure it with a fixed list of customer questions.
Key takeaways
- Definition: GEO is the work that gets your store named and linked in AI answers. For its own AI features, Google calls it SEO.
- Access: A site that blocks OAI-SearchBot is not shown in ChatGPT search answers, according to OpenAI.
- Data: Google, ChatGPT and Perplexity all compare products on price, availability and reviews, from your pages and your product feed.
- Mentions: In an Ahrefs study of 75,000 brands, mentions on other sites tracked AI Overview visibility about three times more closely than backlinks.
- Measurement: AI answers change from run to run, so count the share of answers that name your store over many runs.
- Bottom line: Fix crawler access and product data first, then page content and mentions, and measure with the same questions every month.
Generative engine optimization for ecommerce
Generative engine optimization (GEO) is the work that gets a store named and linked inside AI answers. Answer engine optimization (AEO) is another name for the same goal. For Google, it is still SEO. Its own guide says that optimizing for generative AI search is optimizing for the search experience, so the pages, data and mentions that help in search results also help in AI answers.
The term comes from a 2023 research paper that was accepted at the KDD 2024 conference. Its authors tested content changes on a benchmark of 10,000 search queries. The best changes raised a source’s visibility in AI answers by up to 40%.
The work matters because people click less when an AI summary appears. Pew Research Center tracked the browsing of 900 US adults in March 2025. On Google pages with an AI summary, people clicked a regular result in 8% of visits. Without a summary, they clicked one in 15% of visits. Links inside the summary got a click in 1% of visits.
Product sources in ChatGPT, Gemini, Perplexity and Google AI
Each AI search engine builds a shopping answer from three kinds of source: a search index, product data from stores, and what other websites say. Each company describes its own mix in its documentation. The table sums up what each one states, and what a store can change on its side.
The reach is large. Google says AI Overviews have more than 2.5 billion monthly users and AI Mode more than 1 billion. AI Overviews run in more than 200 countries and territories. They reached Spain in March 2025, and AI Mode followed in October 2025. Perplexity’s Instant Buy, its checkout inside the answer, is available only in the US.
| AI search engine | Main sources | What the store controls |
|---|---|---|
| Google AI Overviews and AI Mode | Google’s index, with several related searches for each question, and the Shopping Graph of more than 50 billion product listings | Indexed pages that can show a snippet, a current Merchant Center feed, structured data that matches the page |
| Gemini app | Content from the Google Search index, used to ground its answers | The Google-Extended rule in robots.txt, plus the same pages and feed as for Google |
| ChatGPT | Its own search crawler OAI-SearchBot, product feeds from merchants, and reviews from public websites | Access for OAI-SearchBot, a product feed (Shopify and Etsy catalogues are already connected), reviews on other sites |
| Perplexity | Pages crawled by PerplexityBot, with sources ranked by authority and relevance, and products by relevance and ratings | Access for PerplexityBot, complete product details, ratings and reviews |
Robots.txt rules for AI crawlers
An AI search engine can only name a store it can read. Open yourstore.com/robots.txt and check three crawlers: OAI-SearchBot for ChatGPT search, PerplexityBot for Perplexity, and Googlebot for Google Search and its AI features. Then check the settings that sit outside robots.txt: the Search Console control, your firewall and your CDN.
Each company documents its crawlers on its own page. The rules below come from those pages.
- OAI-SearchBot: allow it. OpenAI says sites that block it are not shown in ChatGPT search answers. A change takes about 24 hours to apply.
- GPTBot: allow or block it as you choose. It collects content for model training, and OpenAI says each crawler setting is independent of the others.
- PerplexityBot: allow it. Perplexity uses it to show and link websites in its answers, and not to train AI models.
- Google-Extended: this token controls Gemini training and grounding in the Gemini app. Google says it does not change inclusion or ranking in Google Search.
- Search Console: open Settings > Search generative AI. Inclusion is the default. An excluded site gets no traffic or impressions from AI Overviews or AI Mode.
- Firewall and CDN: OpenAI publishes the IP addresses of each crawler. Allow them in your firewall and CDN as well.
Product structured data and product feeds for AI search
AI search engines compare products on facts: price, availability, shipping, returns and ratings. Give each engine the same facts in two places. Put Product structured data on every product page, and send a product feed to Google Merchant Center and to ChatGPT. Google says pages and feed together give a product the widest eligibility in its results.
Schema.org defines the vocabulary. A Product carries a brand, a GTIN, reviews and an aggregate rating. Its Offer carries the price, availability, shipping details and the return policy. Google adds one rule for its AI features: the structured data must match the visible text on the page.
Google’s AI guide says Merchant Center data helps your products appear in AI responses and in other search results. In January 2026, Google announced dozens of new Merchant Center attributes for AI Mode and Gemini, such as answers to common product questions, compatible accessories and substitutes. They go to a small group of retailers first.
For ChatGPT, Shopify and Etsy catalogues are already connected. Other stores apply through OpenAI’s merchant form with a feed that meets its specification. OpenAI says it ranks merchants on availability, price, quality and whether the store makes or mainly sells the item. On PrestaShop, the theme writes the product markup, so check what yours outputs. Our PrestaShop SEO and catalogue work starts with that check.
- Price and availability: the same numbers in the markup, the feed and the page.
- Shipping and returns: shippingDetails and hasMerchantReturnPolicy inside the Offer.
- Identifiers: brand and GTIN on every product that has one.
- Reviews: aggregateRating and review only for reviews that the page shows.
- Variants: a ProductGroup for sizes and colours, with each variant linked to it by isVariantOf.
Category and product page content for AI answers
AI answers quote pages that answer the exact question a shopper asks. Google says AI Overviews and AI Mode issue several related searches across subtopics and data sources for one question. A category or product page needs plain facts in text and short answers to the questions people ask before they pay.
Take the search “waterproof walking boots for wide feet under €150”. It splits into width, waterproofing and price. One product page can answer all three in text, with the price in the markup. Do not make a separate page for every variation of the question. Google says pages made mainly to manipulate rankings or AI responses break its spam policies.
In the GEO paper, three changes worked best: citing sources, adding quotations and adding statistics. Keyword stuffing gave little or no improvement.
- Category pages: open with two or three sentences on what the category holds, who it suits and how to choose. Add a short comparison table.
- Product pages: write sizes, materials, compatibility, delivery times and return terms in text, and not only in images.
- Customer questions: collect them from support emails, chat logs, reviews and Search Console queries. Answer each one on the page where it belongs, in the first sentence.
- Proof: add real numbers with a source for each, and short quotes from experts or buyers.
Brand mentions, reviews and YouTube for AI visibility
AI search engines also read what other websites say about a store: reviews, buying guides, press articles, forums and videos. In the studies available, these mentions track AI visibility more closely than backlinks. That makes reviews and independent coverage part of GEO work, next to your own pages.
Ahrefs studied 75,000 brands. Mentions on other sites had a correlation of 0.664 with AI Overview visibility, where 1 is a perfect match. Backlinks had 0.218. A December 2025 Ahrefs study of ChatGPT, AI Mode and AI Overviews found the strongest link for YouTube mentions, at about 0.737. Correlation does not prove cause, but it shows where to look.
ChatGPT writes review summaries from reviews on public websites, and OpenAI does not verify those reviews. What buyers write about you on other sites shapes how ChatGPT describes your products.
- Ask buyers for reviews on Google and on the review sites your market uses.
- Offer products to independent buying guides and comparison articles in your category.
- Publish short product videos on YouTube with clear titles and descriptions.
- Skip paid or fake mentions. Google’s AI guide warns against inauthentic mentions.
AI visibility tracking for online stores
Measure GEO with a fixed list of customer questions that you run in each AI search engine every month. Count the share of answers that name your store, the share that link to it, and the competitors named instead. Keep the same questions every month, so a change in the numbers means a change in visibility.
One run proves little, because AI answers change from one run to the next. In a January 2026 SparkToro study, 600 volunteers ran 12 prompts 2,961 times in ChatGPT, Claude and Google’s AI. For ChatGPT and Google’s AI, the chance of getting the same list of brands twice was under 1 in 100. The author suggests at least 60 to 100 runs per prompt before you take the average.
Platform reports add a second view. The Generative AI performance report in Search Console shows how often links to your site appeared in AI Overviews and AI Mode. Since 31 August 2026 it is open to all websites. Bing Webmaster Tools launched AI Performance as a public preview in February 2026. It shows how often Copilot and Bing’s AI summaries cite your pages.
Our SEO and AI search service uses the same method: one fixed question list, the same engines, every month.
- Write 20 to 50 questions in your customers’ words: category, comparison, problem and brand questions.
- Run each question many times in ChatGPT, Gemini, Perplexity, Copilot and Google AI Overviews, and use the average.
- Record whether your store is named, whether it is linked, and which competitors appear.
- Compare the result with the Search Console and Bing reports once a month.
GEO checks on our own sites
We run these checks on two sites we know from the inside. RNJ Transport is a road freight forwarder in Lithuania and our client. CreatorFlow is the Instagram DM automation company that our co-founder Vytas runs. Both sites allow named AI crawlers in robots.txt, publish an llms.txt file and serve a Markdown copy of every page.
The RNJ Transport website is built in 11 languages. Its robots.txt names and allows 11 AI and search crawlers, and its pages carry schema such as Organization, FAQPage and Article. We have no AI citation numbers for it yet. The first Search Console review is planned after 18 November 2026.
CreatorFlow is our co-founder’s own SaaS company, not a client. Its robots.txt names 18 AI crawlers with the same rules as all other crawlers, and a check script fails if the two groups differ. From 14 October to 10 November 2026, Search Console counted 676,313 impressions in AI features, 23.5% of its web impressions. Many changes ran at the same time, so we cannot credit that number to one fix.
Common questions
What is the difference between GEO, AEO and SEO?
SEO aims at rankings and clicks in search results. GEO and AEO aim at mentions and links inside AI answers. The base is the same: pages that crawlers can read, clear facts, structured data and mentions on other sites. Google’s own guide says that optimizing for its AI features is still SEO.
Can a store pay to appear in AI answers?
Not inside the answer itself. OpenAI says ChatGPT chooses product results independently and that they are not ads. Perplexity says advertisers cannot pay to list their products. Google says ranked lists in its AI features are not sponsored. What you can change is your data, your pages and what other sites say about you.
Should I block GPTBot to protect my content?
Only if you do not want OpenAI to train models on it. Blocking GPTBot does not affect ChatGPT search. OAI-SearchBot is the crawler for ChatGPT search, and OpenAI says each setting is independent. You can block GPTBot and still allow OAI-SearchBot.
Does my online store need an llms.txt file?
Not for Google. Google says you need no AI text files, markup or Markdown to appear in AI Overviews or AI Mode. We still publish an llms.txt on the sites we build, because it is a small file. We do not count it as the reason for any result.
Does GEO work for a store outside the US?
Yes for Google. AI Overviews run in more than 200 countries and territories, and AI Mode reached Spain in October 2025. Perplexity’s Instant Buy checkout is the exception, because it is available only to users in the US. Crawler access, product data and mentions count in every market.
How do I know if ChatGPT names my store?
Ask ChatGPT the questions your customers ask, with web search on, many times each. Count how often it names and links your store, and which competitors it names instead. Repeat the same questions every month. Before a discovery call, our agents run this check for 20 customer questions in ChatGPT, Gemini and Google.
Sources and checks (28)
- Aggarwal and others, “GEO: Generative Engine Optimization”, KDD 2024 (arXiv)
- Google Search Central: AI features and your website
- Google Search Central: Google’s guide to optimizing for generative AI features on Google Search
- Google: New opportunities, control and insights for website owners (June 2026, updated August 2026)
- Search Console Help: Generative AI performance report (checked November 2026)
- Search Console Help: Search generative AI control
- Google Search Central: Intro to Product structured data
- Google: Google’s common crawlers (Google-Extended)
- Google Shopping Help: Sources of shopping info
- Google: Sundar Pichai’s remarks at the 2026 National Retail Federation (Shopping Graph)
- Google: New tech and tools for retailers to succeed in an agentic shopping era (January 2026)
- Google: AI Overviews expand to over 200 countries and territories (May 2025)
- Google: AI Overviews come to more countries in Europe, Spain included (March 2025)
- Google Spain: AI Mode arrives in Spain (October 2025, in Spanish)
- Schema.org: Product
- Schema.org: Offer (shippingDetails, hasMerchantReturnPolicy)
- OpenAI: Overview of OpenAI crawlers
- OpenAI Help Center: Shopping with ChatGPT Search (checked November 2026)
- ChatGPT: Power product discovery in ChatGPT (for merchants)
- Perplexity: Perplexity crawlers
- Perplexity Help Center: What is Instant Buy? (checked November 2026)
- Pew Research Center: Google users are less likely to click on links when an AI summary appears (July 2025)
- Ahrefs: An analysis of AI Overview brand visibility factors, 75,000 brands
- Ahrefs: Top brand visibility factors in ChatGPT, AI Mode and AI Overviews (December 2025)
- SparkToro: AIs are highly inconsistent when recommending brands or products (January 2026)
- Bing Webmaster Blog: Introducing AI Performance in Bing Webmaster Tools (February 2026)
- RNJ Transport: robots.txt
- CreatorFlow: robots.txt