GEO, AEO and AI Search: A Glossary of 12 Terms for 2026
Twelve AI search terms from agency pitches, each defined in one sentence and checked against Google, OpenAI, Anthropic, Perplexity and Microsoft docs, with what it means for a business site and what to do, if anything.
Key takeaways
- GEO and AEO are two names for one goal: being the source an AI answer names and links. Google says optimizing for its AI features is still SEO, with no extra technical requirements.
- AI companies run separate bots for search, user requests and training. Per their docs, blocking GPTBot or ClaudeBot limits training without removing a site from ChatGPT or Claude search, and Google-Extended does not affect inclusion in Google Search, which includes AI Overviews and AI Mode.
- llms.txt is a 2024 proposal, not a standard. Google says it will neither harm nor help a site in Search, and we found no AI search engine that says it reads the file to choose sources.
- E-E-A-T is not a ranking factor by itself, per Google. It is what Google's quality raters assess, and trust counts most.
- No official metric covers AI visibility across engines. Search Console reports impressions in AI Overviews and AI Mode; Bing Webmaster Tools reports citations and citation share for Copilot and Bing's AI answers.
GEO, AEO and AI search are new names for an old job: when a buyer asks a question, the answer should name your business and link to your site. Most AI search pitches use the same twelve terms. Each entry below gives a one-sentence definition, why it matters for a business site, what to do (sometimes nothing yet) and the primary source.
These are also the terms of our SEO & GEO work, so you can check any proposal against them, ours included.
| Term | In one line |
|---|---|
| GEO | Getting named and linked in AI answers |
| AEO | Another name for the same goal |
| AI Overviews, AI Mode | Google's AI answers inside Search |
| RAG, grounding | Answers written from retrieved pages |
| Query fan-out | One question split into many searches |
| AI crawlers | Bots for search, user requests and training |
| llms.txt | A proposed summary file for AI tools; no effect on Google Search |
| Structured data | Page facts in schema.org vocabulary, usually JSON-LD |
| Entity | A thing a search engine can tell apart from others |
| E-E-A-T | What Google's raters assess, trust first |
| Zero-click search | A search that ends without a click |
| AI visibility, citation share | How often answers name or cite you |
What do GEO and AEO mean?
GEO and AEO are two labels for the same goal: being the source an AI answer names and links. Google treats both as part of SEO, so neither needs a separate playbook.
GEO (generative engine optimization)
GEO is the practice of making your content more likely to be used and cited when an AI tool such as ChatGPT, Perplexity or Google's AI Overviews writes an answer.
Why it matters. The term comes from a 2023 research paper, accepted at KDD 2024, that introduced GEO as a way for content creators to raise their visibility in generative engine responses. If your page is not among the sources an answer cites, the buyer does not see your name.
What to do. Treat GEO as SEO plus crawler access. Our guide to getting recommended by ChatGPT and Google AI checks the common GEO tactics against the evidence.
AEO (answer engine optimization)
AEO is the practice of shaping pages so that an answer engine, such as a search feature, voice assistant or AI chat, gives your content as the answer.
Why it matters. Google's guide to generative AI search, updated July 2026, names both AEO and GEO and warns that many suggested hacks "aren't effective or supported by how Google Search actually works". Its conclusion: optimizing for generative AI search "is optimizing for the search experience, and thus still SEO."
What to do. If a proposal sells SEO, GEO and AEO as three separate services, ask for the task list of each. Most lines will repeat.
What are AI Overviews and AI Mode?
AI Overviews are AI-written summaries in Google results that help people "get to the gist of a complicated topic or question more quickly"; AI Mode is a separate search mode for questions that need "further exploration, reasoning, or complex comparisons". Both answer from Google's own index, and Google says there are no additional technical requirements to appear in them: a page must be indexed and eligible to show with a snippet.
Why it matters. Google launched AI Overviews in 2024, opened AI Mode to everyone in the US in May 2025 and expanded it to more than 180 countries and territories in English that August. AI Mode is still small: in SparkToro's 2026 study, only 0.34% of US Google searches reached it from January to April 2026.
What to do. Nothing beyond being indexed and allowed to show a snippet. To keep a page or a passage out, use the snippet controls: nosnippet, data-nosnippet, max-snippet or noindex.
How do AI answers find your pages?
They search before they write: the model retrieves pages from a search index, writes the answer from them and links the pages it used.
RAG and grounding
Retrieval-augmented generation (RAG) is a method in which an AI model first retrieves relevant documents and then writes its answer from them; Google calls the same idea grounding.
Why it matters. The method was named in a 2020 paper by Patrick Lewis and colleagues, presented at NeurIPS 2020. Google's AI guide says RAG, "also known as grounding", uses its core ranking systems to retrieve "relevant, up-to-date web pages from our Search index." A page that is not in the index cannot be retrieved, so it cannot be cited.
What to do. Check in Google Search Console and Bing Webmaster Tools that your key pages are indexed. Fix indexing before anything labelled AI.
Query fan-out
Query fan-out is when an AI system splits one question into several related searches, runs them at once and combines what it finds into one answer.
Why it matters. Google's AI guide defines it as "a set of concurrent, related queries generated by the model", and Google says AI Mode breaks a question down into subtopics. So the pages an answer cites can come from searches the buyer never typed. A question like "which CRM suits a 20-person agency" may also set off searches for pricing, integrations and reviews.
What to do. List the follow-up questions your buyers ask, such as price ranges, timelines, comparisons and risks, and answer each in the first lines of a page. Our GEO guide shows how fan-out changes which pages get cited.
Which AI crawlers visit your site, and how does robots.txt control them?
An AI crawler is a bot an AI company uses to fetch web pages for a search index, for a user's request or for model training. Most companies run a separate bot for each job, and each bot reads its own group in robots.txt, so you can allow one job and block another.
Which bot does which job, from each company's documentation, checked October 2026.
| Job | OpenAI | Anthropic | Perplexity | |
|---|---|---|---|---|
| Search index for answers | OAI-SearchBot | Claude-SearchBot | PerplexityBot | Googlebot (Search, including AI Overviews and AI Mode) |
| Fetching a page a user asked for | ChatGPT-User | Claude-User | Perplexity-User | Google-Agent and other user-triggered fetchers |
| Training | GPTBot | ClaudeBot | None listed: PerplexityBot is not used for training | Google-Extended, a robots.txt token rather than a bot |
What blocking does, per the docs:
- OpenAI. Each setting is independent: you can block GPTBot and still allow OAI-SearchBot. Opting out of OAI-SearchBot keeps a site out of ChatGPT search answers, apart from navigational links. For ChatGPT-User, robots.txt may not apply, since a person asked for the page.
- Anthropic. Its bots honor robots.txt. Blocking ClaudeBot asks that future content stay out of training; blocking Claude-SearchBot "may reduce your site's visibility" in Claude's search results.
- Perplexity. PerplexityBot surfaces and links sites in Perplexity search and is not used to crawl content for AI foundation models. Perplexity-User, sent when a user asks, "generally ignores robots.txt rules".
- Google. Google-Extended decides whether Google may use your content to train Gemini models and to ground answers in Gemini Apps and Vertex AI; it "does not impact a site's inclusion in Google Search". AI Overviews and AI Mode are part of Search, so they follow Googlebot and the snippet controls. Google's user-triggered fetchers, such as Google-Agent, generally ignore robots.txt.
What to do. Allow the search bots of every AI tool your buyers use, both in robots.txt and in your CDN or firewall, which can block bots that robots.txt allows. Treat the training bots as a business decision. Per the docs, blocking GPTBot or ClaudeBot does not take you out of ChatGPT or Claude search, and blocking Google-Extended leaves Google Search alone but keeps your pages out of grounding in the Gemini app.
What is llms.txt, and does any AI engine use it?
llms.txt is a proposed Markdown file at the root of a website, /llms.txt, that gives AI tools a short summary of the site and links to its key pages. It is a proposal, not a standard, and we found no AI search engine that says it reads the file to choose sources.
Why it matters. Jeremy Howard proposed it in September 2024 "to help agents use a website"; the proposal reached its second version in August 2026. Who says what:
- Google: the AI guide says the file "will neither harm nor help" a site's visibility or rankings in Google Search.
- Chrome: Lighthouse checks llms.txt, flags only server errors and calls the file "optional at the moment".
- OpenAI, Anthropic and Perplexity: their crawler pages, linked above, cover robots.txt and say nothing about reading llms.txt.
What to do. Nothing yet, unless the basics are done. We publish one on nextgensoft.co because it costs nothing, and we expect no citations from it. On the new site for Encore, a Dubai PR, media and events agency, llms.txt shipped next to schema and Open Graph. The bigger fix was 34 pages, each with one H1 and its own meta description, where the old site had 8 pages, no meta descriptions and 38 H1 tags on the homepage. Our audit checklist covers when llms.txt is worth checking at all.
What are structured data, schema.org and JSON-LD?
Structured data is, in Google's words, "a standardized format for providing information about a page and classifying the page content". Schema.org is the vocabulary most of it uses, JSON-LD is the format Google recommends, and none of it is required to appear in AI answers.
Why it matters. Google's introduction to structured data recommends JSON-LD as "the easiest solution for website owners to implement and maintain at scale". Schema.org was founded by Google, Microsoft, Yahoo and Yandex. Markup can make a page eligible for rich results in regular search, but Google's AI guide is explicit: "there's no special schema.org markup you need to add" for generative AI search.
What to do. Mark up the facts you already show, such as your organization, products, reviews, articles and breadcrumbs, and validate them in Google's Rich Results Test. Skip offers of special "AI schema". On Gastro.News, an Austrian restaurant guide, venue pages carry schema.org Restaurant, AggregateRating and Review markup in JSON-LD, and every state and district hub carries ItemList and BreadcrumbList. The programmatic SEO case study shows how the guide grew to 2,600+ pages without spam.
What do entities and E-E-A-T mean for a business site?
An entity is your business as a search engine understands it: one thing with one name and one set of facts. E-E-A-T is the yardstick Google's human raters use to judge whether results deserve trust. Neither is a setting you switch on.
Entity
An entity is a real-world thing, such as a company, person, place or product, that a search engine can tell apart from others with a similar name.
Why it matters. Google built its Knowledge Graph around entities in 2012 and summed up the shift as "things, not strings". An answer about your company can only be as accurate as the engine's picture of which company you are. Google says Organization markup can help it understand your administrative details and "disambiguate your organization in search results".
What to do. Use the same company name, address and phone number everywhere. Add Organization markup to your home page, with sameAs links to your official profiles. Keep your Google Business Profile current: Google's AI guide says Business Profiles and Merchant Center can help your products and services show up in AI responses.
E-E-A-T
E-E-A-T stands for experience, expertise, authoritativeness and trustworthiness: the qualities Google's quality raters assess when they check whether Search results are good.
Why it matters. Google is direct: "E-E-A-T itself isn't a specific ranking factor", raters "have no control over how pages rank", and of the four aspects "trust is most important". The current rater guidelines, dated September 2025, tell raters to look at what a site says about itself, what independent sources say about it and what is visible on the page.
What to do. Don't buy "E-E-A-T optimization" as a ranking fix. Do what a careful buyer checks anyway: a real About page, named authors with relevant experience, case studies with facts that can be checked and reviews on independent platforms.
How do you measure results when searches end without a click?
Count where your business appears, not only how many people visit. Most Google searches already end without a click, and the official AI reports measure impressions and citations, not traffic.
Zero-click search
A zero-click search is a search that ends without a click on any result.
Why it matters. In SparkToro's 2026 study, based on Similarweb's US panel for January to April 2026, 68.01% of Google searches ended without a click. The author ties the rise mainly to the growth of AI Overviews. A buyer can read your name in an answer and contact you without ever visiting your site.
What to do. Judge search work by more than visits: add "AI assistant" to the "How did you hear about us?" field on your forms, and watch branded searches in Search Console.
AI visibility and citation share
AI visibility is how often AI answers name or link your business for the questions your buyers ask; citation share is your part of all the citations an answer engine shows for a query.
Why it matters. There is no official metric across engines, and Google's AI guide notes: "No third-party tool has access to our internal ranking or AI systems." Two engines report on their own answers:
- Google: Search Console's Generative AI performance report shows how often links to your site appeared in AI Overviews and AI Mode. Google rolled it out to all sites on August 31, 2026.
- Bing: Bing Webmaster Tools' AI Performance report, in public preview since February 2026, counts citations in Copilot, Bing's AI summaries and select partners. In June 2026 Bing added Citation Share: for a grounding query, the percentage of all citations shown that go to your site. Bing calls it "not a ranking system or a competitive scoreboard".
We have found no comparable report for ChatGPT, Claude or Perplexity.
What to do. Record a baseline this month: both reports, plus a fixed panel of 20 to 30 buyer questions asked in ChatGPT, Perplexity and Google AI Mode in a fresh session, noting who gets named and linked. Repeat it monthly.
Where should you start this week?
Start with three checks that take an afternoon and need no budget.
- Read your robots.txt at yoursite.com/robots.txt, and your CDN's bot settings. Confirm OAI-SearchBot, Claude-SearchBot, PerplexityBot and Googlebot are not blocked, then decide on GPTBot, ClaudeBot and Google-Extended on purpose.
- Open the two reports, Generative AI performance in Search Console and AI Performance in Bing Webmaster Tools, and write down this month's numbers. Not verified in Bing yet? Do that first.
- Ask five buyer questions in ChatGPT, Perplexity and Google AI Mode, each in a fresh session. Note who gets named, who gets linked and whether the facts about you are right.
For the technical side in one report, our free SEO & GEO check looks at your home page and up to nine more pages the way Google, Bing and AI crawlers see them, including AI crawler access, Bing indexing and llms.txt.
When the checks turn up more than your team wants to fix, our SEO & GEO service covers the full list: technical SEO, crawler access, pages for the questions your buyers ask, markup and monthly AI visibility tracking. To plan a budget, see what SEO costs in 2026.
Questions we get asked
What is the difference between GEO, AEO and SEO?
GEO (generative engine optimization) and AEO (answer engine optimization) both mean getting your business named and linked when AI tools answer a question. Google's guide to generative AI search calls this still SEO: AI Overviews and AI Mode answer from Google's index, with no extra technical requirements. The work is the same: crawler access, pages that answer the exact question and what other sites say about you.
Should I block GPTBot and other AI crawlers?
Decide by what each bot does. GPTBot and ClaudeBot collect training data; OAI-SearchBot, Claude-SearchBot and PerplexityBot decide whether you can appear in ChatGPT, Claude and Perplexity search answers. OpenAI says each setting is independent, so you can block training and still allow search. Google-Extended covers Gemini training and grounding in the Gemini apps, and has no effect on inclusion or ranking in Google Search.
Do I need an llms.txt file?
No. llms.txt is a 2024 proposal for a Markdown file that summarizes a site for AI tools. Google says it will neither harm nor help a site in Search, and Chrome's Lighthouse calls it optional. We found no AI search engine that says it reads the file to pick sources. It costs little to publish, so add it after the basics if you like, but don't expect citations from it.
Is E-E-A-T a ranking factor?
Not by itself. Google says E-E-A-T itself isn't a specific ranking factor; its systems use a mix of signals that tend to match content with good experience, expertise, authoritativeness and trust. Its quality raters judge E-E-A-T to check whether results are good, but their ratings don't change how any page ranks. Of the four, Google says trust matters most.
How do I measure my visibility in AI answers?
Use the official reports where they exist: Search Console's Generative AI performance report for impressions in AI Overviews and AI Mode, and Bing Webmaster Tools' AI Performance for citations in Copilot and Bing's AI answers. For ChatGPT, Claude and Perplexity, ask a fixed set of buyer questions every month in a fresh session and log who gets named and linked.
What share of Google searches end without a click?
In SparkToro's 2026 study, based on Similarweb's US panel for January to April 2026, 68.01% of Google searches ended without a click. The author ties the rise mainly to the growth of AI Overviews. For a business site, that means fewer visits per search, so being the source an answer names matters more than it did.
Sources
- Optimizing your website for generative AI features on Google Search · Google Search Central, 2026
- AI features and your website · Google Search Central, 2025
- AI in Search: Going beyond information to intelligence · Google, 2025
- AI Mode in Search gets new agentic features and expands globally · Google, 2025
- GEO: Generative Engine Optimization · arXiv (KDD 2024), 2024
- Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks · arXiv (NeurIPS 2020), 2020
- Overview of OpenAI Crawlers · OpenAI, 2026
- Does Anthropic crawl data from the web, and how can site owners block the crawler? · Anthropic, 2026
- Perplexity Crawlers · Perplexity, 2026
- Google's common crawlers · Google Search Central, 2026
- Google's user-triggered fetchers · Google Search Central, 2026
- The /llms.txt file · llmstxt.org (Jeremy Howard), 2024
- llms.txt audit · Chrome for Developers, 2026
- Introduction to structured data markup in Google Search · Google Search Central, 2025
- About Schema.org · Schema.org, 2026
- Introducing the Knowledge Graph: things, not strings · Google, 2012
- Organization structured data · Google Search Central, 2026
- Creating helpful, reliable, people-first content · Google Search Central, 2026
- Search Quality Rater Guidelines · Google, 2025
- In 2026, Less than One Third of Google Searches Still Send a Click · SparkToro, 2026
- Generative AI performance report · Search Console Help, 2026
- Introducing AI Performance in Bing Webmaster Tools Public Preview · Microsoft Bing, 2026
- New AI Visibility Insights in Bing Webmaster Tools: Intents, Topics, Citation Share, Compare · Microsoft Bing, 2026
Share this article
About the author
Artsiom Yankovich
Head of Design & Strategy
Leads design and strategy, from product logic to brand. 9+ years, 150+ projects.
