Key SEO and AI Search Terms, Explained

GEO (Generative Engine Optimisation) is the practice of structuring web content so that AI-powered search engines — including Google AI Overviews, ChatGPT, Perplexity, and Gemini — cite your pages in their generated answers rather than only linking to them. Unlike traditional SEO, which optimises for blue-link rankings, GEO focuses on the structured signals, answer-shaped prose, and schema markup that AI grounding systems use to extract citable facts. OrbiRank's audit pipeline includes an AI-crawler accessibility check — verifying that GPTBot, ClaudeBot, and PerplexityBot are allowed to crawl the site — as a prerequisite for GEO eligibility.

An AI Overview is Google's AI-generated summary answer displayed above organic search results, drawing from crawled web pages and structured data to produce a synthesised response. Pages that appear as sources in AI Overviews typically have clear structured data markup, well-organised headings, and factual, answer-shaped content that Google's grounding model can extract with confidence. Being cited in an AI Overview can drive qualified traffic independently of a page's blue-link ranking position.

Core Web Vitals are Google's set of user-experience performance metrics used as ranking signals: Largest Contentful Paint (LCP, target under 2.5 seconds), Interaction to Next Paint (INP, target under 200 milliseconds), and Cumulative Layout Shift (CLS, target under 0.1). Google incorporates Core Web Vitals scores into its Page Experience ranking signal, meaning poor scores can suppress a page's position even when its content is relevant and authoritative. OrbiRank measures LCP, INP, CLS, and TTFB for both mobile and desktop via the PageSpeed Insights API as part of every site audit.

INP (Interaction to Next Paint) is Google's Core Web Vitals responsiveness metric, introduced as a stable signal in 2024 to replace First Input Delay (FID). It measures the latency of every user interaction on a page — clicks, taps, and keyboard inputs — and reports the worst interaction delay observed during the visit, with a passing threshold of under 200 milliseconds. Because INP captures the full range of interactions rather than only the first one, it is a more accurate indicator of a page's runtime responsiveness under real usage conditions.

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness — Google's quality rater framework for evaluating whether content meets a high standard for the people who will read it. Concrete on-page E-E-A-T signals include author bylines with verifiable credentials, datePublished and dateModified schema markup, outbound citations to authoritative sources, and Organisation schema on the homepage. E-E-A-T is also a key eligibility signal for AI Overview citation: AI grounding models favour content that demonstrates clear authorship, factual sourcing, and editorial accountability.

Keyword cannibalization occurs when two or more pages on the same domain compete for the same primary keyword, splitting ranking signals and preventing either page from ranking as strongly as a single authoritative page would. Search engines receive conflicting relevance signals when multiple URLs target the same query, and may alternate which URL they rank — causing rank instability and suppressing overall visibility. Resolving cannibalization typically requires consolidating competing pages, applying canonical tags to designate the preferred URL, or differentiating the pages' target keywords through content restructuring.

Backlink velocity is the rate at which a domain acquires new referring domains over a given period, typically measured per week or per month. A natural, gradually accelerating velocity — consistent with growing content and word-of-mouth — is a positive quality signal; unnaturally fast acquisition (hundreds of new referring domains in days) can trigger Google's SpamBrain spam detection and result in a manual or algorithmic penalty. Monitoring velocity alongside referring domain quality helps distinguish sustainable link growth from manipulative patterns.

llms.txt is an emerging web standard — modelled on robots.txt — hosted at a domain root (e.g. https://example.com/llms.txt) that provides AI agents and large language models with a structured, machine-readable index of the site's key pages, purpose, and capabilities. The format uses plain text with Markdown-style headings and bullet links so that AI assistants performing web-grounded retrieval can efficiently identify the most important pages without crawling the entire site. Publishing a well-maintained llms.txt is an AEO best practice because it gives AI crawlers a prioritised entry point, reducing the risk that the most commercially important pages are overlooked during indexing.

E-E-A-T signals are the concrete on-page elements that demonstrate Experience, Expertise, Authoritativeness, and Trustworthiness to both human quality raters and AI grounding systems. Key signals include: author bylines with professional credentials, datePublished and dateModified schema markup on articles, outbound citations linking to peer-reviewed or authoritative sources, Organisation schema on the homepage, and an accessible About or Team page describing who is responsible for the content. Strengthening E-E-A-T signals is particularly important for YMYL (Your Money or Your Life) topics — finance, health, legal — where Google's quality raters apply the highest scrutiny.

A canonical tag is an HTML link element — <link rel="canonical" href="..."> — placed in the <head> of a page to tell search engines which URL is the preferred, authoritative version when duplicate or near-duplicate content exists across multiple URLs. It is commonly used to consolidate ranking signals from URL variants caused by query parameters (e.g. ?sort=price), session IDs, or syndicated content that appears on more than one domain. Incorrectly self-referencing canonical tags, missing canonical tags on paginated content, or canonicalling to a noindex URL are among the most frequently found technical SEO issues in site audits.

A redirect chain is a sequence of two or more HTTP redirects that occur between an original URL and its final destination — for example, URL A → URL B → URL C. Each hop in the chain adds network latency for real users and can dilute the link equity (PageRank) passed from inbound backlinks, because some signal is lost at each redirect step. Search engines recommend collapsing chains to a single direct redirect wherever possible; OrbiRank's technical audit pipeline includes automated redirect chain detection and reports the full hop sequence for affected URLs.

Structured data is machine-readable markup embedded in a web page — typically as a JSON-LD script tag — using the Schema.org vocabulary to help search engines and AI systems understand factual properties of the page such as content type, author, price, rating, product name, and other entities. JSON-LD (JavaScript Object Notation for Linked Data) is Google's recommended format for structured data because it can be placed anywhere in the document, is easy to validate, and does not require modifying the visible HTML. Accurate structured data is a prerequisite for rich results (star ratings, FAQ dropdowns, product prices in SERPs) and a key signal for AI Overview citation eligibility.

AI crawler accessibility refers to whether a site's robots.txt file permits or blocks the dedicated crawlers used by AI platforms to index content for their answer engines: GPTBot (OpenAI/ChatGPT), ClaudeBot (Anthropic/Claude.ai), and PerplexityBot (Perplexity AI). A site that blocks these crawlers with a Disallow directive becomes invisible to AI-generated answers regardless of its content quality — the AI simply has no data to cite. OrbiRank audits robots.txt for AI crawler access as part of every technical SEO scan and flags any blocking rule as a GEO (Generative Engine Optimisation) risk.

Topical authority is a measure of how comprehensively a domain covers a subject area, assessed by the breadth and depth of its interlinked content addressing the topic's full range of subtopics and related questions. A site with high topical authority on, say, technical SEO will have well-structured content covering audits, canonicalisation, redirects, structured data, Core Web Vitals, and related subtopics — all internally linked in a coherent topic cluster — rather than isolated articles on individual keywords. Search engines and AI grounding models both favour topically authoritative domains as citation sources because they signal sustained, expert-level coverage rather than opportunistic keyword targeting.

MCP (Model Context Protocol) is an open standard introduced by Anthropic that allows AI assistants — including Claude, ChatGPT plugins, Cursor, Windsurf, and Gemini CLI — to connect to external data sources and tools via a standardised JSON-RPC interface. An MCP server exposes a tools/list endpoint that AI clients query to discover available capabilities; the client then invokes specific tools by name, passing structured parameters and receiving structured responses without writing custom integration code for each AI platform. OrbiRank's MCP server is live at /mcp and exposes 10 tools — including run_audit, get_rank_positions, get_content_brief, and generate_content — accessible to any MCP-compatible AI client.