The 36 terms that define AI search

AEO is the practice of structuring web content so AI answer engines (ChatGPT, Perplexity, Google AI Overviews, Claude) cite your page as the source of their answer.

llms.txt is a plain-Markdown file at /llms.txt that declares your site's authoritative content to AI crawlers. Proposed by Jeremy Howard, 2024.

An AI citation is the inline source attribution that ChatGPT, Perplexity, or Google AI Overviews displays alongside its synthesized answer. Citations link back to source pages.

GEO is the practice of structuring content so AI engines retrieve and include it in the context window used to generate their answer — one layer earlier than AEO.

GPTBot is OpenAI's web crawler that gathers training data for future GPT models. Distinct from OAI-SearchBot (search index) and ChatGPT-User (on-demand browse).

PerplexityBot is the crawler that builds Perplexity AI's own search index. Must be explicitly allowed in robots.txt to appear in Perplexity answer citations.

ClaudeBot is Anthropic's web crawler. Family includes ClaudeBot, Claude-Web, Claude-SearchBot, Claude-User, and anthropic-ai for different surfaces of the Claude product.

Google AI Overviews is the AI-generated answer panel at the top of certain Google search results. Inherits classic Google rank directly.

SGE was Google's experimental AI search interface launched in 2023 and renamed to AI Overviews in 2024-05. The two terms refer to the same product family.

Google-Extended is Google's crawler for Gemini training data and non-search AI features. Separate from Googlebot — blocking it does not affect Google Search ranking.

Schema.org is the shared structured-data vocabulary used by Google, Bing, ChatGPT, Perplexity, and most major search and AI engines for entity and content markup.

FAQPage is the Schema.org type for marking up question-and-answer content. Pages with FAQPage schema are cited 2-3x more often for question queries by AI engines.

Structured data is machine-readable metadata embedded in web pages — typically Schema.org vocabulary in JSON-LD format — that helps search and AI engines extract facts.

Core Web Vitals are three Google-defined metrics — Largest Contentful Paint, Interaction to Next Paint, and Cumulative Layout Shift — that measure real-user page experience.

TTFB is the elapsed time between a user's HTTP request and the first byte of the server's response. Under 800ms is good; under 200ms is excellent.

A canonical URL declares which version of a page is the authoritative one when duplicate or near-duplicate URLs exist. Set via the rel='canonical' link tag.

robots.txt is a plain-text file at /robots.txt that tells web crawlers which paths on a site they may or may not crawl. AI bots respect it.

Open Graph is the meta-tag protocol developed by Facebook for controlling how a page appears when shared on social media (Facebook, LinkedIn, Twitter, Slack, Discord).

Sitemap.xml is an XML file listing every indexable URL on a site. Helps search engines discover content efficiently. Should be linked from robots.txt.

noindex is a meta robots directive that tells search engines not to include a page in their index. Used for thin pages, search-results pages, and per-user surfaces.

A 301 redirect is a permanent server-side redirect from one URL to another. Passes ~90-100% of link equity to the new URL. Use for permanent URL changes.

LCP is a Core Web Vitals metric — the time to render the largest visible content (hero image, headline, video) above the fold. Under 2.5 seconds is good.

CLS is a Core Web Vitals metric measuring visual stability during page load. Under 0.1 is good. Caused by images without dimensions, ads, and late-loading fonts.

INP is the newest Core Web Vitals metric (replaced FID in 2024). Measures responsiveness — the delay between a user input and the next visual update. Under 200ms is good.

HTTPS is HTTP over TLS — encrypts traffic between browser and server. Required for SEO ranking, AEO citation candidacy, and modern browser features.

rel="noopener" is a link attribute that prevents the destination page from accessing the parent window via window.opener. Security best practice for target="_blank" links.

E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is Google's quality signal framework used by both classic search ranking and AI Overviews source selection.

The Google Knowledge Graph is Google's structured database of entities (people, places, organizations, products). Pages tied to a recognized entity cite more reliably in AI Overviews.

Internal linking is the practice of linking between pages on the same site. Distributes PageRank, helps crawlers discover content, and signals topical authority.

Backlinks are inbound links from other domains. Still the strongest classic Google ranking signal. Modestly weighted by AI engines for source-quality assessment.

Anchor text is the clickable text in a hyperlink. Tells search engines what the destination page is about; the second-most-important on-page signal after the title tag.

The title tag is the <title> element in the page <head>. Displayed as the clickable SERP heading. The single strongest on-page SEO signal.

rel="nofollow" is a link attribute telling search engines not to pass link equity to the destination. Common on user-generated content, paid links, and untrusted outbound links.

The meta description is a 140-160 character summary in the page head that Google may display as the SERP snippet. Not a ranking factor; high CTR impact.

The H1 is the primary heading of a page, marked with <h1>. One H1 per page is best practice. Should mirror the title tag's primary keyword.

Alt text is the alternative text attribute on <img> elements. Describes the image for screen readers and search engines. Critical for accessibility and image SEO.