This glossary covers every key term in AI visibility: GEO, AI Share of Voice, Visibility Score, LLM, RAG, JSON-LD, prompts, citations, entities, and more. Bookmark it as your reference for understanding how brands are discovered in AI-generated answers.
A – C
| Term | Definition |
|---|---|
| AI Citation | When an AI model explicitly names, links to, or attributes information to a specific brand or URL in its generated response. Citations can be direct (URL link) or indirect (brand name mention). Higher citation quality means higher Visibility Score. |
| AI Overviews | Google's feature that displays AI-generated summaries at the top of search results. Brands that appear in AI Overviews gain significant visibility without users clicking through to a website. A key target of GEO strategies. |
| AI Share of Voice (SOV) | The percentage of AI-generated responses that mention your brand for a set of prompts, compared to competitors. Formula: SOV = (Your mentions ÷ Total mentions) × 100. A SOV above 30% is considered strong. Full guide → |
| Content Cluster | A group of interlinked articles covering multiple facets of a single topic. Content clusters build topical authority that AI models reward with higher citation rates. Each article links to the others with descriptive anchor text. |
E – G
| Term | Definition |
|---|---|
| Entity | A clearly defined concept (brand, person, product, topic) that AI models can recognize and distinguish. Entity clarity — consistent naming, descriptions, and structured data — is a core GEO signal. The clearer your entity, the more likely AI models cite you accurately. |
| Entity Recognition | The process by which an AI model identifies and classifies entities (brands, products, people) within text. Rankio uses entity recognition to detect both direct and contextual brand mentions in AI responses. |
| GEO (Generative Engine Optimization) | The practice of optimizing content so AI models (ChatGPT, Gemini, Perplexity) cite and recommend your brand. Combines structured data, entity clarity, topical authority, and citation-friendly formatting. GEO is to AI search what SEO is to Google. Full guide → |
J – L
| Term | Definition |
|---|---|
| JSON-LD | JavaScript Object Notation for Linked Data — a structured data format used to describe entities (Organization, Product, Article, FAQPage) in a way search engines and AI models can parse. Adding JSON-LD is one of the fastest GEO wins, often producing results within days. |
| LLM (Large Language Model) | An AI model trained on massive text datasets that generates human-like text. Examples: GPT-4 (OpenAI), Gemini (Google), Claude (Anthropic). LLMs are the "engines" behind AI search — the systems that decide which brands to mention in their answers. |
| llm.txt | A text file placed at the root of a website (similar to robots.txt) that provides AI models with structured information about the site's purpose, key pages, and how to cite it. Helps LLMs understand and accurately represent your brand. |
P – R
| Term | Definition |
|---|---|
| Parametric Knowledge | Information embedded in an LLM's weights during training. The model "knows" this information without needing to fetch it live. Brands can influence parametric knowledge through consistent, authoritative web presence over time. |
| Prompt | The question or instruction a user sends to an AI model. In GEO, prompts are the queries your audience types into ChatGPT, Gemini, or Perplexity. Your content must answer these prompts to earn citations. |
| Prompt Set | A curated collection of 30–100+ prompts representing how your target audience interacts with AI. Used for Visibility Score analysis and SOV testing. Should include discovery, comparison, branded, and intent prompts. |
| RAG (Retrieval-Augmented Generation) | A technique where an AI model fetches live web pages before generating an answer, combining retrieved context with its parametric knowledge. Perplexity and ChatGPT with browsing use RAG. Your pages must be retrievable and well-structured to be selected by the RAG pipeline. |
| Recommendation Strength | A metric measuring whether an AI model merely mentions your brand or actively recommends it. "We recommend X" scores higher than "X is an option". Part of Rankio's Visibility Score (15% weight). |
S – V
| Term | Definition |
|---|---|
| Sentiment Analysis | Determining whether an AI model's mention of your brand is positive, neutral, or negative. Positive mentions score 1.0 in Rankio's Visibility Score; neutral 0.5; negative 0.1. Even negative mentions contribute to visibility, but minimally. |
| Structured Data | Machine-readable information added to web pages (typically as JSON-LD) that helps search engines and AI models understand content. Common schemas: Organization, Product, Article, FAQPage, BreadcrumbList, SoftwareApplication. |
| Topical Authority | The depth and breadth of a website's coverage on a specific subject. AI models prefer citing sources that demonstrate comprehensive expertise. Built through content clusters, internal linking, and consistent entity definitions. |
| Visibility Score | Rankio's composite metric (0–100) that aggregates 30+ weighted signals to quantify how visible your brand is in AI-generated answers. Components: presence (25%), citation quality (20%), position (15%), recommendation (15%), sentiment (10%), consistency (10%), frequency (5%). Full methodology → |
Ready to measure your AI visibility?
Run your first analysis and discover your Visibility Score across ChatGPT, Gemini, and Perplexity.