Skip to content

GEO, AEO, LLMO and AI SEO: a plain glossary of AI search terms

What GEO, AEO, LLMO, AIO and AI SEO mean, how they overlap, and plain definitions of crawler, grounding, RAG, query fan-out and citation for AI search.

By Crawlito · 7 min read

In short

  • GEO, AEO, LLMO, AIO and AI SEO are overlapping marketing labels for being visible in AI-generated answers; no standards body defines any of them.
  • GEO has the clearest origin: a 2023 research paper titled GEO: Generative Engine Optimization by Aggarwal and colleagues.
  • AIO is ambiguous: it is used both for AI optimization and as shorthand for Google's AI Overviews, so define it before using it in a report.
  • Google describes grounding as providing content from the Google Search index to a model at prompt time, and says AI Overviews and AI Mode may use query fan-out, issuing several related searches.
  • Training crawlers, search crawlers and user-triggered fetchers are the three crawler roles most AI operators document, each with its own user-agent token.
On this page

GEO (generative engine optimization), AEO (answer engine optimization), LLMO (large language model optimization), AIO (AI optimization) and AI SEO are different names for roughly the same goal: making your content findable, readable and quotable in answers produced by AI tools. None of them is defined by a standards body, and the people who use them often mean slightly different things. Google's position is that optimizing for generative AI search is still SEO.

The labels matter less than the vocabulary underneath them. Words like crawler, grounding, RAG, query fan-out and citation describe how AI search actually works, and they show up in operator documentation, vendor reports and audit findings. This glossary defines both sets in plain terms, with sources where a definition comes from an operator.

The optimization labels

TermStands forWhat people usually mean
SEOSearch engine optimizationBeing crawled, indexed and ranked by search engines
AEOAnswer engine optimizationBeing the source an engine uses for a direct answer; older than generative AI, from featured snippets and voice assistants
GEOGenerative engine optimizationBeing visible inside answers written by generative AI
LLMOLarge language model optimizationBeing understood, mentioned or recommended by LLM-based tools
AIOAI optimization, or AI OverviewsEither the general goal, or Google's AI Overviews feature specifically
AI SEOAI search engine optimizationAn umbrella term for SEO work aimed at AI search

A few notes on where these come from:

  • GEO traces to a research paper, GEO: Generative Engine Optimization by Pranjal Aggarwal and colleagues, first posted in November 2023. It tested content changes against generative engines on a research benchmark. It is a useful study; it is not a description of how any particular commercial assistant picks sources today.
  • AEO predates large language models. It grew up around featured snippets and voice assistants that read out a single answer.
  • AIO is the one to be careful with. In Google-focused writing it often means AI Overviews, the feature. In agency proposals it often means AI optimization, the practice. If a report says "AIO traffic", ask which.

Google addresses the labels directly in its AI optimization guide, which says that from Google Search's point of view, optimizing for generative AI search is optimizing for the search experience, and points readers to its guidance on evaluating SEO advice. Our post on GEO versus SEO covers what genuinely differs and what is ordinary technical SEO. We use a narrower term, AI search readiness, for the technical question of whether AI tools can reach and read your pages at all.

AI search products

TermWhat it is
AI OverviewsAI-generated summaries shown on some Google Search results pages, with links to supporting pages
AI ModeA Google Search mode for longer, conversational queries, with follow-up questions
ChatGPT searchOpenAI's web search inside ChatGPT, which cites sources
PerplexityAn answer engine that searches the web and cites sources for each answer
Microsoft CopilotMicrosoft's assistant; Microsoft says it grounds web answers through the Bing search service
Claude web searchAnthropic's web search inside Claude

Google's AI features page says that to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to show in Google Search with a snippet, with no additional technical requirements.

Crawler vocabulary

Crawler. A program that fetches web pages automatically, usually following links or a list of URLs, to build an index or a dataset.

Fetcher. A program that retrieves a specific URL on request rather than crawling. Google and several AI operators use this word for user-triggered tools.

User agent and token. The user agent is the name a client sends in the User-Agent header. The token is the short name used in robots.txt, such as GPTBot. They are usually, but not always, the same word.

Training crawler. A crawler whose content may be used to train future models, such as GPTBot, ClaudeBot and CCBot.

Search crawler. A crawler that builds the index an assistant searches when it answers with sources, such as OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot and Bingbot.

User-triggered fetcher. An agent that fetches a page because a person asked, such as ChatGPT-User, Claude-User and Perplexity-User. Some operators say robots.txt may not apply to these.

Opt-out token. A robots.txt name that does not belong to a separate crawler but controls how content is used. Google-Extended and Applebot-Extended are the main examples.

Verified bot. A request confirmed to come from the operator it claims, usually by reverse DNS or published IP ranges. Anyone can send a user-agent string; verification checks the network.

The three crawler roles are covered in detail in training, search and user-triggered crawlers.

How AI answers are produced

Large language model (LLM). A model trained on large amounts of text to predict and generate language. ChatGPT, Claude, Gemini and others are built on LLMs.

Training data. The text a model learned from. Content in training data can shape what a model says, but the model does not normally cite it as a source.

Knowledge cutoff. The date after which a model's training data has no information. Search features exist partly to answer questions about anything newer.

Retrieval. Looking up documents, for example from a search index, and passing them to the model while it answers.

RAG (retrieval-augmented generation). The pattern of retrieving documents and giving them to a model so its answer is based on them. The term comes from a 2020 research paper, Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks by Lewis and colleagues. Most AI search features are a form of RAG.

Grounding. Tying a model's answer to specific source content. Google's crawler documentation describes grounding as providing content from the Google Search index to the model at prompt time to improve factuality and relevancy.

Index. The stored, searchable copy of pages a search system has crawled. If a page is not in the index an assistant searches, that assistant has no copy of it to retrieve.

Query fan-out. Google says AI Overviews and AI Mode may issue multiple related searches across subtopics and data sources to build one answer. A page can be chosen for a sub-query the user never typed. See query fan-out explained.

Hallucination. A confident statement from a model that is not supported by its sources or is simply false.

Measurement vocabulary

Citation or source link. A link to your page shown alongside an AI answer as a source.

Mention. Your brand or product named in an answer, with or without a link.

Referral traffic. Visits that arrive by clicking a link in an AI tool. Some tools pass a referrer; some visits arrive without one and look like direct traffic.

Zero-click. A search where the user gets an answer without visiting any site.

AI visibility, share of voice. Vendor metrics that track how often a brand appears for a set of prompts. They are not standardized, and results depend on the prompts, location and timing used. AI citations versus referral traffic explains why the two numbers rarely match.

Files and signals you will hear about

TermWhat it isStatus
robots.txtRules for crawlers by user-agent token and pathStandardized as RFC 9309; the documented control for AI crawlers
llms.txtA proposed Markdown file listing a site's key pages for LLMsA proposal; see our llms.txt guide for who documents reading it
Structured dataMachine-readable facts in the page, usually JSON-LD using schema.orgUsed by search engines for rich results
Content signalsLines in robots.txt stating preferences for search, AI input and trainingIntroduced by Cloudflare
TDMRepA protocol for reserving text and data mining rightsW3C Community Group report, not a W3C standard
nosnippet, max-snippetRobots meta rules that limit snippetsGoogle says they also limit use in AI Overviews and AI Mode

Frequently asked questions

Is GEO different from SEO?

Mostly in emphasis. The technical groundwork is the same: crawl access, indexable pages, clear structure and text in the HTML. Google says optimizing for generative AI search is optimizing for the search experience, which is SEO.

What is the difference between AEO and GEO?

AEO is older and focused on direct answers such as featured snippets and voice results. GEO is focused on answers written by generative AI. In practice the terms are used interchangeably.

Grounding means basing a model's answer on specific retrieved content rather than only on what the model learned in training. Google defines it as providing content from the Google Search index to the model at prompt time.

Does AIO mean AI Overviews?

Sometimes. It is used both for Google's AI Overviews and for AI optimization in general, so check which one a report or proposal means.

Start with whether AI tools can read your site

Whatever label you use, nothing else works if crawlers cannot fetch and read your pages. The free check tests a page for crawler access, raw HTML content and structure, and an audit does the same across your site with prioritized fixes.

Filed under Strategy and audits. Topics: ai-search, strategy, geo, ai-crawlers