Skip to main content
Get early access

Free to join. Just your email.

GEO and AI visibility glossary

AI visibility and Generative Engine Optimization come with a lot of new terms. Below are clear explanations of the most important ones, from GEO and prompt tracker to citations, share of voice and llms.txt. New to the topic? Start with what is GEO. The practical steps to make your site readable for AI are in how AI crawlers read your site.

GEO (Generative Engine Optimization)
Optimizing your website and content so that generative AI search engines such as ChatGPT, Gemini and Perplexity mention and recommend your business as a source in their answers.
AEO (Answer Engine Optimization)
Optimizing to appear as the answer in answer engines. In practice it largely overlaps with GEO; both are about being mentioned in AI answers.
LLM (Large Language Model)
A large language model that generates text based on patterns in training data, such as the models behind ChatGPT, Gemini and Claude.
Prompt
The question or instruction a user gives an AI. In AI visibility, the prompt replaces the keyword as the unit of measurement.
Prompt tracker
A tool that automatically tests prompts in AI platforms and measures whether and how your brand appears in the answers. See the GreatCompany prompt tracker.
AI Overviews
The AI-generated answers at the top of Google that summarize sources instead of only showing links.
Mention
An AI answer in which your business or brand is named, with or without a link. How often that happens for your customers' questions is the basis of measuring AI visibility.
Citation
A source reference or link an AI shows with an answer. Getting cited more often increases your visibility and traffic.
Share of voice
The share of AI answers in your market in which your brand appears, compared with competitors.
Visibility score
One number that summarizes how often and how prominently you are mentioned in AI answers, so you can follow it over time.
Hallucination
An incorrect or made-up answer from an AI. Hallucinations can arise about your brand; you want to monitor and correct them.
llms.txt
A text file in the root of your site that aims to help AI systems find your most important content. Its effect on AI search is still limited and evolving.
Structured data (schema)
Structured markup (JSON-LD) that helps search engines and AI understand your content, such as Product, FAQ and Article.
AI crawler
A bot from an AI platform that fetches your pages to read content and use it in answers, such as GPTBot and PerplexityBot.
Grounding
Backing up an AI answer with real sources from the web, instead of relying on the model alone.
RAG (Retrieval-Augmented Generation)
A technique where an AI first retrieves relevant sources and uses them to generate an answer. It explains why good, findable content matters.
Agentic commerce
Shopping where an AI agent not only recommends but also helps complete the purchase, often through product cards in the AI interface.
Sentiment
The tone in which an AI talks about your brand, from positive to negative. Part of brand monitoring in AI.
Prompt-first content strategy
Creating content from the questions customers ask AI, instead of from loose keywords.
Question cluster
A group of related questions around one intent. In GEO you optimize per cluster instead of per keyword.
E-E-A-T
Experience, Expertise, Authoritativeness, Trustworthiness. Signals of reliability that affect whether AI uses you as a source.
SSR (Server-Side Rendering)
The HTML of a page is generated on the server, so crawlers and AI see the content right away without running JavaScript.
AI Mode (Google)
Google's conversational search mode, where you ask follow-up questions and get AI answers instead of a list of links. Different from AI Overviews: that is a summary above the regular results, AI Mode is a full conversation.
Attribution
Assigning a lead or revenue to a channel. Harder in GEO than in SEO, because an AI answer often brings no click but does influence the purchase.
Bot blocking
Keeping crawlers out through robots.txt, a firewall or IP filtering. A trap in GEO: block AI crawlers too strictly and you become invisible in AI answers.
Chunking
Splitting a page into smaller pieces that an AI can retrieve and cite separately. Clear headings and well-defined paragraphs increase the chance the right piece is found.
Citation rate
The percentage of relevant questions where your site is named as a source. One of the most important numbers in GEO.
Content authority
How reliable and authoritative AI considers your content on a topic. You build it with consistency, factuality and mentions elsewhere.
Direct answer
An answer the user gets right away without clicking through. In GEO you want your information to form that answer, including a source reference.
Embedding
A mathematical representation of text as a series of numbers, which AI uses to compare meaning. It explains why AI also finds content that does not literally contain your keyword.
Entity
A recognizable concept, person or business in the eyes of AI. Once you are recognized as an entity, AI can mention you consistently. See also entity disambiguation.
Entity disambiguation
Determining which entity is meant by a name with several meanings. Important if your brand name is also a common word or another company.
Fine-tuning
Further training an existing AI model on specific data. Not something you steer as a business, but useful to know in conversations about AI.
Foundation model
A large, pre-trained model on which applications are built, such as GPT, Claude and Gemini. The foundation under the AI search engines you want to be visible in.
Generative engine
A search engine that generates answers instead of only showing links. ChatGPT, Perplexity, Gemini and AI Overviews are generative engines. The G in GEO.
GPTBot
The OpenAI crawler that collects content for training models. Blocking it reduces your chance of being mentioned in ChatGPT.
OAI-SearchBot
The OpenAI crawler for ChatGPT Search. Unlike GPTBot (training), this bot decides whether you appear in current search answers.
Inference
The moment a model processes a question and generates an answer. Every prompt is an inference.
Information gain
The unique value your content adds on top of what already exists. AI prefers to cite a source that says something new over yet another repetition.
Knowledge Graph
Google's structured database of entities and their relationships. Appearing in it strengthens how recognizable you are to AI.
Knowledge Panel
The information box next to Google's search results about a recognized entity. A sign that you are recorded as an entity.
Long-tail query
A specific, longer question with less volume but much higher intent, such as which accounting software fits a contractor who is self-employed. Exactly where AI search excels.
Machine-readable content
Content that not only people but also machines can read in a structured way: schema markup, clear headings and plain facts.
Multimodal
A model that processes more than text: images, audio and video too. Relevant because you want your images and videos to be findable as well.
NLP (natural language processing)
The field that lets computers understand human language. The underlying technology that lets you ask a question instead of typing a keyword.
Offer object
Structured data with price, availability and conditions, which AI agents use to show and sell your product. The basis for agentic commerce.
Perplexity
An AI search engine that gives answers with a visible source for every claim. Because of that transparency it is easy to measure and influence.
Prompt appearance
Measures which prompts you appear in the answer for. The GEO equivalent of keyword rankings.
GEO ranking factors
The factors that decide whether AI picks your content. Unlike SEO, factuality, structure, freshness and mentions elsewhere weigh more than backlinks.
Retrieval
Fetching relevant sources as the first step of an AI answer. If you are not retrieved here, you cannot be cited either.
SGE (Search Generative Experience)
The old name for Google's AI search feature, now AI Overviews. You still come across it in older articles.
Token
The unit in which a model processes text, roughly a word or part of a word. Among other things it determines how much context a model can take in.
Training data
The text a model was trained on. It determines what the model knows on its own. GEO focuses mainly on what the model retrieves live, not on training data.
Vector database
A database that stores and searches embeddings. The memory behind RAG systems.
WAF (web application firewall)
A security layer for your site. Set too strictly, it blocks AI crawlers and makes you invisible by accident.

From terms to results

Measure with the prompt tracker whether AI mentions your business.

Further reading

See whether ChatGPT mentions your business.

Join the waitlist. When GreatCompany opens in the US, you get your first report right away: who ChatGPT recommends in your market and what to improve on your site.

Free to join. When we open in the US, your free AI visibility report is first in line.

Curious what AI crawlers see on your site? About the technical scanPro $49, you do it yourself · Or automate your SEO for $99, built for WordPress

Opening in the US

Be first when GreatCompany opens.

Leave your email and website. When we open, you get your free AI visibility report first: which customer questions ChatGPT mentions you for, who it names instead, and what to fix first.

Free. One email when we open, nothing else. By joining you accept our Terms and Privacy policy.