The AI era multiplied search jargon fast. We defined the 20 terms you will meet most, in marketer language and without hype.
The practice of getting your brand mentioned in the answers of generative engines like ChatGPT and Gemini. SEO aims to climb the results list. GEO aims to enter the answer itself. GEO (Generative Engine Optimization) explained →
Preparing content for surfaces that answer questions directly: AI summaries, voice assistants, chat engines. It overlaps heavily with GEO. The two usually describe the same work. AEO (Answer Engine Optimization) explained →
AI visibility is how often and in what way your brand appears in AI engine answers. It is a measurable metric: in which questions are you mentioned, at which position, with which framing? AI visibility explained →
The model type behind ChatGPT, Gemini and Claude. Trained on text, it produces probabilistic answers, which is why the same question does not always get the identical answer. LLM (Large Language Model) explained →
Prompt tracking is asking the questions your customers ask to the engines on a schedule and recording the brands in the answers. This is the core method of AI visibility measurement. Prompt tracking explained →
The visibility score is signals like mention rate, position and share of voice reduced to a single 0-100 number. How rankzupAI computes it is written plainly on the methodology page. Visibility score explained →
Mention rate is the share of tracked runs where your brand appears in the answer. Mentioned in 8 of 20 runs means a 40 percent rate. Mention rate explained →
Position is where your brand comes in the list when an answer names multiple brands. First position in an AI answer is the equivalent of the first result in classic search. Position explained →
A citation is the source link an engine bases its answer on. Perplexity and AI Overviews show citations openly. The sites in that list shape the answer's content. Citation explained →
Grounding is a model basing its answer on current web sources. Gemini's search-backed answers are the classic example. Grounding makes the sources it reads measurable. Grounding explained →
AI Overviews is the AI summary at the top of Google search, compiled from multiple sources. A large share of users read only this summary before deciding. AI Overviews explained →
Google search turned into a conversation: follow-up after follow-up, often without ever opening the results list. Available in Turkish in Türkiye since 18 February 2026. Google AI Mode explained →
Your share of all brand mentions across the answers. The most honest summary of the competitive picture: what percent of the market conversation is yours? Share of voice explained →
Sentiment is the framing an engine uses when mentioning you: positive, neutral or negative. Being mentioned is not enough. How you are mentioned moves the purchase decision. Sentiment explained →
Hallucination is a model stating something that does not exist as if it did. Wrong prices, invented features or false claims about your brand fall here. Watching for them is reputation work. Hallucination explained →
The user getting the answer on the search page or inside the assistant and never visiting a site. AI answer surfaces amplify this behavior. Visibility is now won before the click. Zero-click search explained →
RAG is the model fetching information from external sources before answering and building the answer on it. Perplexity and every search-backed answer works on this architecture. RAG (Retrieval-Augmented Generation) explained →
Crawlability is whether AI bots like GPTBot, Google-Extended and PerplexityBot can read your site. A misconfigured robots.txt means the engines never read you, and your chance of entering answers closes at the door. AI bots and crawlability explained →
A confidence band shows a low-data score as a range instead of a single number. LLM answers are probabilistic. Honest measurement shows the uncertainty instead of hiding it. Confidence band explained →
A question without any brand name in it: 'best protein powder' is brand-free, 'is X trustworthy' is branded. Fair competitor comparison uses brand-free questions only. Branded ones inflate their owner artificially. Brand-free query explained →
AI searchability is how readable and quotable your site is for AI engines. It differs from classic crawlability in its goal: not getting indexed, but becoming the source an answer is built from. Bot access, clear answer structure and extractable content decide it. AI searchability explained →
These generative engine optimization terms come up once you move from measuring visibility to fixing it.
GEO differs from SEO in how success is judged, while both rest on crawlable pages with clear answers. SEO is judged by rank in a results list. GEO is judged by whether the answer names your brand. The full comparison is in GEO vs SEO.
llms.txt is a Markdown file at a site's root that summarizes the site for AI agents and links to its key pages. It is still a proposal at llmstxt.org, not a standard every engine reads. See the llms.txt guide for setup.
Structured data is markup that labels facts on a page, such as a product's price or an article's author. Most sites use the schema.org vocabulary, usually written as JSON-LD.
Chunking is splitting a page into passages so a retrieval system can match each passage to a question. A section that stands on its own survives chunking better than one that leans on earlier paragraphs.
Query fan-out is an engine issuing several related searches across subtopics to build one answer. Google says both AI Overviews and AI Mode may use this technique.
See your own numbers on every metric in this glossary with a single scan, and learn where your brand stands.
Published: Last updated: