AI systems now decide which businesses get recommended, cited, and trusted – and they do it in a language of their own. This glossary defines the terms that matter for understanding how AI visibility works, in plain language, with a link to the full guide for each term. The definitions are short on purpose: each one gives you enough to understand the term, and the linked post explains it properly.
The terms
- AEO (Answer Engine Optimization) – Making your business easy for AI answer engines to find, understand, and recommend. What is Answer Engine Optimization (AEO)?
- AI answer – The generated reply an AI system gives to a question, built from the sources it retrieved. What is an AI Answer?
- AI knowledge base – The collection of information an AI system draws on about a business or a topic, assembled from many sources. What is an AI Knowledge Base?
- AI visibility – How often, and how accurately, AI systems mention, cite, or recommend your business in their answers. What is AI Visibility?
- Embeddings – Numbers that represent the meaning of text, so an AI system can tell which pieces of content are about the same thing. What are Embeddings in AI?
- Generative AI – AI that creates new text, images, or other content, rather than only classifying or retrieving what already exists. What is Generative AI?
- GEO (Generative Engine Optimization) – Making your business easy for generative AI engines to cite and recommend in their answers. What is Generative Engine Optimization (GEO)?
- Hallucination – When an AI system produces a confident answer that is false or made up, because it has no reliable source for the fact. What is AI Hallucination?
- Knowledge base poisoning – Corrupting the information sources an AI system draws on, so its answers about a business or topic become wrong. What is Knowledge Base Poisoning?
- Large language model (LLM) – An AI model trained on large amounts of text that can understand and generate human language, and that powers chatbots and AI answers. What is a Large Language Model (LLM)?
- Machine learning – The method by which AI systems learn patterns from data, rather than being programmed with explicit rules. What is Machine Learning?
- Model poisoning – Tampering with an AI model’s training so it produces wrong or biased results. What is Model Poisoning?
- Prompt injection – Hiding instructions in content an AI system reads, so it follows the hidden instruction instead of the user’s request. What is Prompt Injection?
- Retrieval-augmented generation (RAG) – A technique where an AI system retrieves relevant documents before answering, so its reply is grounded in those sources. What is Retrieval-Augmented Generation (RAG)?
- RAG poisoning – Corrupting the documents a retrieval-augmented system retrieves, so its answers are based on false information. What is RAG Poisoning?
- Training data poisoning – Injecting false or misleading data into the data an AI model is trained on, so the model learns the wrong thing. What is Training Data Poisoning?
- Vector database – A database that stores embeddings, so an AI system can find content by meaning rather than by exact words. What is a Vector Database?
Terms change as the technology changes, and so does this glossary. If a term is missing, or a definition is out of date, tell us through the contact page and we will correct it.