Glossary

Search and AI search terms

How Google, Bing and AI answer engines find, read and cite pages: crawling, indexing, structured data, GEO and the files that steer AI crawlers.

AEO vs GEO: AEO (answer engine optimization) focuses on being chosen as the direct answer to a question, as in featured snippets, voice answers and AI Overviews. GEO (generative engine optimization) focuses on being mentioned, cited and recommended in AI-generated answers more broadly. They overlap heavily, and many people use the terms interchangeably.
AI citation: An AI citation is a source link an AI assistant or AI search feature shows alongside its answer, pointing to the page it drew from. Perplexity, ChatGPT search, Google AI Overviews and Copilot all cite sources. Being cited sends visitors and signals that the engine trusted your page for that answer.
AI crawler: An AI crawler is a bot that fetches web pages for an AI company, either to collect training data, to build a search index for AI answers, or to fetch a page live when a user asks. Examples include GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and CCBot. Each can be allowed or blocked in robots.txt.
AI Mode: AI Mode is Google Search's conversational AI experience, launched in the US in 2025, where users ask complex questions and get a generated answer with links, then follow up. It uses query fan-out to run many related searches at once. Its traffic appears in Search Console with regular web search.
AI Overviews: AI Overviews are AI-generated summaries Google shows at the top of some search results, with links to the web pages they drew from. They launched broadly in the US in May 2024 after testing as the Search Generative Experience (SGE). Google says pages need no special markup to appear; normal SEO eligibility applies.
AI share of voice: AI share of voice is the share of AI-generated answers in your category that mention your brand, compared with competitors, across a fixed set of prompts. If your product appears in 12 of 50 answers and your top competitor in 30, your share is lower. It's the AI-search version of share of search.
AI visibility: AI visibility is how often and how favourably a brand appears in answers from AI assistants and AI search: whether it's mentioned, cited with a link, recommended, and described correctly. It's measured by running a fixed set of buyer prompts across engines over time, since there's no equivalent of a ranking report.
Answer engine optimization: Answer engine optimization (AEO) is optimising content to be picked as the direct answer to a question by search features and AI assistants: featured snippets, voice assistants, AI Overviews and chatbots. It focuses on short, clear, self-contained answers placed where machines can extract them.
Average position: Average position in Google Search Console is the mean ranking of your topmost result across all impressions for a query or page. Position 1 is the top. Because it averages across searches, locations and devices, it's a trend indicator, not a fixed rank. Lower numbers are better.
Bing Webmaster Tools: Bing Webmaster Tools is Microsoft's free tool for site owners to see how their site performs in Bing: search performance, indexing, sitemaps, backlinks, SEO reports and IndexNow submissions. Bing's index also feeds other search and AI products, so its data matters beyond bing.com.
Brand mention: A brand mention is any reference to your company or product name on another website, social platform or forum, with or without a link. Mentions build awareness and help search engines and AI systems understand your brand as an entity. For AI answers, being mentioned in the right places matters as much as being linked.
Canonical tag: A canonical tag is a link element, rel="canonical", in a page's head that tells search engines which URL is the main version when the same or similar content is available at several URLs. Google treats it as a strong hint, alongside redirects, internal links and sitemaps, when choosing the canonical.
ClaudeBot: ClaudeBot is Anthropic's web crawler that collects public pages that may be used to train Claude models. Anthropic also runs Claude-SearchBot, which indexes pages to improve Claude's search answers, and Claude-User, which fetches pages when a Claude user asks. All respect robots.txt.
Click-through rate: Click-through rate (CTR) is the share of people who clicked your result out of everyone who saw it: clicks divided by impressions. In Search Console it's reported per query and page. CTR depends on position, what else is on the results page, and how compelling your title and description are.
Content cluster: A content cluster is a group of pages on one subtopic, linked to each other and to a central pillar page that covers the topic broadly. Each cluster page answers one specific question in depth. The linking helps search engines see the coverage and helps readers move between related answers.
Content decay: Content decay is the gradual decline in a page's search traffic and rankings over months, usually because the content went out of date, competitors published better pages, search intent shifted, or AI answers now cover the question. Refreshing decayed pages is often faster than writing new ones.
Crawl budget: Crawl budget is the number of URLs a search engine can and wants to crawl on a site in a given time. It depends on how fast and healthy the server is (crawl capacity) and how much the engine wants the content (crawl demand). Google says it mainly matters for very large sites.
E-E-A-T: E-E-A-T stands for Experience, Expertise, Authoritativeness and Trustworthiness, the qualities Google's search quality rater guidelines ask human raters to assess in content and its creators. Trust is the most important. It's not a single ranking factor, but Google's systems aim to reward content that shows these qualities.
Entity: In SEO, an entity is a distinct, identifiable thing, such as a company, product, person, place or concept, that search engines and AI systems recognise and store facts about, independent of the exact words used. Becoming a recognised entity helps a product appear correctly in knowledge panels and AI answers.
FAQPage schema: FAQPage schema is schema.org structured data that marks up a list of questions and answers on a page. It once produced expandable FAQ rich results for many sites; since August 2023 Google shows them mainly for well-known government and health sites. The markup is still valid and helps machines read the Q&A.
Featured snippet: A featured snippet is a highlighted answer Google shows near the top of results, extracted from one web page with a link to it. It appears for questions where Google thinks a short excerpt answers the query: a paragraph, list or table. Pages earn it by answering clearly in a well-structured section.
Generative engine optimization: Generative engine optimization (GEO) is the practice of making your brand and pages more likely to be mentioned, cited and recommended in answers from AI systems such as ChatGPT, Perplexity, Gemini, Claude and Google's AI Overviews. It builds on SEO but targets being quoted inside an answer rather than ranked in a list.
GEO vs SEO: SEO (search engine optimization) aims to rank pages in search engine results; GEO (generative engine optimization) aims to get a brand mentioned, cited and recommended inside AI-generated answers. GEO builds on SEO, since AI answers often retrieve top-ranking pages, but adds answer-first writing, sourcing, consistent facts and third-party mentions.
Google Search Console: Google Search Console is Google's free tool for site owners to see how their site performs in Google Search: queries, clicks, impressions and positions, which pages are indexed and why others aren't, sitemaps, links, manual actions and security issues. It's the most direct source of Google search data for your site.
Google-Extended: Google-Extended is a robots.txt token, not a separate crawler, that controls whether content Google crawls can be used to train future Gemini models and for grounding in Gemini apps. Blocking it doesn't affect Google Search rankings or inclusion in AI Overviews, which follow normal Googlebot rules.
GPTBot: GPTBot is OpenAI's web crawler that collects publicly available pages that may be used to train its AI models. It identifies itself with the user agent token GPTBot and respects robots.txt. Blocking GPTBot does not remove a site from ChatGPT search, which uses a separate crawler, OAI-SearchBot.
Grounding: Grounding is connecting an AI model's answer to specific sources, such as search results, documents or databases, so its claims are based on retrieved information rather than only on training data. Grounded answers usually show citations. Google uses the term for Gemini answers grounded in Google Search.
Hallucination: An AI hallucination is a confident but false statement generated by a language model: an invented fact, a wrong number, a fake citation or a feature a product doesn't have. It happens because models generate likely text, not verified facts. Grounding in sources and consistent information online reduce it.
Helpful content: Helpful content is Google's term for content created primarily to benefit people rather than to rank. Google launched a "helpful content system" in 2022 and folded it into its core ranking systems in March 2024. Its guidance asks whether content shows first-hand experience, satisfies the reader and isn't made mainly for search engines.
Impressions: In Google Search Console, an impression is counted each time a link to your site appears in search results a user saw, even without scrolling to it in some layouts or clicking it. Impressions show visibility; clicks show traffic. Rising impressions with flat clicks often means low positions, weak snippets or AI answers.
Indexing: Indexing is when a search engine stores a page it has crawled in its index, the database it searches when someone types a query. Only indexed pages can appear in results. Crawling comes first; after crawling, the engine decides whether the page is useful and unique enough to index.
IndexNow: IndexNow is an open protocol that lets a website notify search engines the moment a URL is added, updated or deleted, instead of waiting to be crawled. Microsoft Bing and Yandex launched it in 2021; Seznam, Naver and Yep also take part. Google doesn't. One submission is shared with all participating engines.
IndexNow vs sitemap: An XML sitemap lists your URLs for search engines to fetch on their own schedule (pull). IndexNow lets you notify participating engines the moment URLs change (push). Google uses sitemaps but not IndexNow; Bing, Yandex and others support both. Most sites should use both.
JSON-LD: JSON-LD (JSON for Linked Data) is a W3C standard format for writing linked data in JSON. On websites it's the most common way to add structured data: a script tag holding a JSON object that uses schema.org types to describe the page. Google recommends it over microdata and RDFa.
Keyword cannibalization: Keyword cannibalization happens when two or more pages on the same site target the same search intent, so they compete with each other and Google alternates between them or ranks neither well. Not every overlap is a problem; it is when the pages answer the same question for the same reader.
Keyword difficulty: Keyword difficulty (KD) is an SEO tool's estimate, usually 0 to 100, of how hard it would be to rank on the first page for a search. Most tools base it largely on the backlink strength of the pages currently ranking. Each tool calculates it differently, so scores aren't comparable across tools.
Knowledge Graph: The Knowledge Graph is Google's database of entities and the facts and relationships between them, introduced in 2012 with the phrase "things, not strings". It powers knowledge panels, the boxes of facts shown for people, companies and products. Getting your company into it makes brand searches show a panel.
Large language model: A large language model (LLM) is an AI model trained on very large amounts of text to predict and generate language. ChatGPT, Claude and Gemini are built on LLMs. They answer questions, write and summarise, and, when connected to search or tools, look things up and take actions.
llms.txt: llms.txt is a proposed plain-text Markdown file at a site's root (/llms.txt) that gives language models a short summary of the site and a curated list of links to its most useful pages. It's an invitation about what to read, not an access rule. Major AI companies haven't committed to reading it.
llms.txt vs robots.txt: robots.txt tells crawlers which paths they may crawl, using allow and disallow rules, and is honoured by major search engines and AI crawlers. llms.txt is a proposed Markdown file that summarises a site and lists its key links for language models; it controls nothing. Use robots.txt for access and llms.txt as an optional guide.
Long-tail keyword: A long-tail keyword is a specific, usually longer search with lower volume, such as "free AI directories with dofollow links" rather than "AI directories". Each gets few searches, but together they make up most search demand, and they're easier to rank for and more likely to convert.
Meta description: A meta description is an HTML meta tag that summarises a page, which search engines may show as the snippet under the title in results. It isn't a ranking factor, but a clear one helps people choose your result. Google often writes its own snippet from the page when that better matches the query.
Noindex: Noindex is an instruction, in a robots meta tag or an X-Robots-Tag HTTP header, that tells search engines not to include a page in their index. The page must stay crawlable so the engine can see the instruction. It's the right tool for keeping pages like logins, thank-you pages or thin pages out of results.
OAI-SearchBot: OAI-SearchBot is OpenAI's crawler for ChatGPT's search features. It finds and indexes pages so ChatGPT can show them as sources and links in answers. OpenAI says it isn't used to train models. Sites that block it won't be shown in ChatGPT search answers, though they may still appear as plain links.
People Also Ask: People Also Ask (PAA) is a Google results box listing related questions that expand to show a short answer from a web page, with a link. Clicking one loads more questions. PAA shows what else people want to know about a topic, which makes it a free source of questions to answer on your pages.
PerplexityBot: PerplexityBot is Perplexity's crawler that indexes web pages so they can appear as sources in Perplexity's answers. Perplexity says it isn't used to train foundation models. A separate agent, Perplexity-User, fetches pages when a user's question requires it. Sites that block PerplexityBot won't be indexed for Perplexity answers.
Pillar page: A pillar page is a broad page that covers a whole topic at a high level and links to detailed cluster pages for each subtopic. It targets the main, broad search for the topic and acts as the hub of a content cluster. Category and hub pages often play this role.
Programmatic SEO: Programmatic SEO is creating many search-targeted pages from a template and a dataset, such as one page per city, integration, comparison or directory. It works when each page carries unique, useful data for its query. Thin pages that only swap a keyword risk being treated as scaled content abuse.
Query fan-out: Query fan-out is a technique AI search uses for complex questions: the system splits the question into many related sub-queries, runs them all and combines the results into one answer. Google has described AI Mode and AI Overviews using it, so pages can be cited for parts of a question they answer well.
Retrieval-augmented generation: Retrieval-augmented generation (RAG) is a way of making a language model answer from specific documents: the system first retrieves relevant passages from a search index or database, then gives them to the model to write the answer. AI search engines use it to answer from the web with citations.
robots.txt: robots.txt is a plain-text file at the root of a website that tells crawlers which paths they may or may not crawl, using User-agent, Allow and Disallow rules. It can also list your sitemap. It controls crawling, not indexing: a blocked page can still appear in results if other sites link to it.
Scaled content abuse: Scaled content abuse is Google's spam policy name for generating many pages primarily to manipulate search rankings rather than help users, whether by AI, automation, people or a mix. Introduced in March 2024, it replaced the narrower "spammy automatically generated content" policy. Sites can be demoted or get manual actions.
Search intent: Search intent is the goal behind a search: what the person wants to know, do or buy. It's usually grouped as informational, navigational, commercial and transactional. Pages rank best when they match the intent Google sees in the current results, in both content and format.
SERP: A SERP (search engine results page) is the page a search engine shows after a query. Today it holds much more than ten blue links: ads, AI Overviews, featured snippets, People Also Ask, videos, images, local results and more. What appears changes by query and affects how many clicks organic results get.
Structured data: Structured data is code added to a page, usually as JSON-LD using the schema.org vocabulary, that describes what the page contains in a machine-readable way: an article, a product, an organisation, an FAQ, a breadcrumb trail. Search engines use it to understand pages and to show rich results.
Title tag: A title tag is the HTML <title> element of a page. Search engines usually use it as the clickable headline in results (Google calls it the title link), and browsers show it in tabs. A good one says what the page is in roughly 50 to 60 characters. Google may rewrite titles it finds unhelpful.
Topical authority: Topical authority is the idea that search engines trust a site more on a subject when it covers that subject thoroughly and well, with many connected pages, rather than one page in isolation. Google has said it values sites with expertise on a topic. It's built with depth, internal links and links from relevant sites.
Topical map: A topical map is a plan of every subtopic, question and page a site needs to cover a subject thoroughly, grouped into clusters with a pillar page each, and showing how the pages will link. It turns keyword research into a content plan aimed at topical authority.
Trailing slash redirect: A trailing slash redirect sends a URL without a final slash to the version with one (or the reverse), so each page has a single address. It's normal, but if your internal links, sitemap or canonical use the other form, every link passes through a redirect and Search Console reports "Page with redirect".
XML sitemap: An XML sitemap is a file that lists the URLs on a site you want search engines to know about, often with the date each was last modified. It helps crawlers find pages, especially new ones and ones with few links. A sitemap suggests pages to crawl; it doesn't guarantee indexing.
Zero-click search: A zero-click search is a search that ends on the results page without a click to any website, because the answer was shown directly: in a featured snippet, knowledge panel, AI Overview, calculator or local pack. A large share of searches end this way, which changes how pages get value from search.

Other glossary sections

Glossary A to Z