Glossary
Definitions of all important terms in AI-search and GEO.
A
- AI-search
- Searching for information through AI-powered tools like ChatGPT, Perplexity, Gemini and Claude, instead of traditional search engines.
- AISO
- AI-Search Optimization: the set of techniques to become discoverable and citable in AI-powered search environments.
- AEO
- Answer Engine Optimization: an older term for optimizing content for direct answers in search engines.
- AI-agents(AI agents)
- Autonomous AI systems that proactively perform tasks: researching, comparing, deciding and taking actions on behalf of users.
- Answer Engine
- A search engine that provides direct answers instead of links to websites. Examples are ChatGPT, Perplexity and Google AI Overviews.
B
- Brand mention
- A mention of your brand name in AI-generated content, with or without a direct link to your website.
C
- Citatie(Citation)
- A mention of your website as source in an AI-generated answer, often with a link.
- ChatGPT
- OpenAI's most popular AI assistant, used by hundreds of millions of people for search, research and advice.
- Citation rate
- The percentage of AI answers that contain a link to your website as source. A key performance indicator for GEO.
- Claude
- AI assistant by Anthropic, known for accuracy, safety and processing long documents.
- Copilot
- Microsoft's AI assistant, integrated into Windows, Office 365, Bing and Edge. Reaches millions of business users.
- Crawl budget
- The amount of resources a bot spends indexing your website. Important for large sites with lots of content.
E
- Entity
- A recognizable unit like a person, company, product or concept that AI systems can identify and connect.
- Embedding
- Numerical representation of text that AI uses to understand meaning and context. Foundation for semantic search.
F
- Few-shot
- AI technique where the model performs a task with a few examples as context, without specific training.
G
- GEO
- Generative Engine Optimization: optimizing your online presence so AI systems understand and cite your brand.
- Gemini
- Google's AI model, integrated into Google Search (AI Overviews), Workspace and Android. Reaches billions of searchers.
- GPTBot
- OpenAI's web crawler that collects content for ChatGPT. Can be allowed or blocked via robots.txt.
- Grounding
- Process of anchoring AI answers in factual, verifiable sources. Reduces hallucinations.
H
- Hallucination
- When an AI model presents incorrect or fabricated information as fact. Risk with poorly documented topics.
K
- Knowledge Graph
- A knowledge structure capturing entities and their relationships, used by search engines and AI to understand connections.
L
- LLM
- Large Language Model: a large language model trained on vast amounts of text, the foundation of AI assistants like ChatGPT.
- llms.txt
- A special text file for AI crawlers with structured information about your company, services and expertise.
M
- Mention
- A mention of your brand or company name in an AI answer, with or without source link.
- Mention rate
- The percentage of AI answers that mention your brand (with or without link). Basic metric for AI visibility.
P
- Prompt testing
- Systematically testing how AI tools respond to specific questions relevant to your brand.
- Perplexity
- AI search engine that generates answers with explicit source citations. Popular with researchers and professionals.
- PerplexityBot
- Perplexity's web crawler that retrieves real-time information for answers. Can be allowed via robots.txt.
R
- RAG
- Retrieval Augmented Generation: technique where AI models retrieve real-time information from external sources to generate accurate answers.
S
- Structured data
- Technical markup (like Schema.org) that helps search engines and AI systems understand your content.
- Semantic search
- Searching based on meaning and context rather than exact keywords. Foundation of modern AI search engines.
T
- Token
- Smallest unit of text that an LLM processes, usually a word or word part. Determines costs and context limits.
V
- Vector database
- Database optimized for storing and searching embeddings. Essential for RAG systems.
Z
- Zero-shot
- AI technique where the model performs a task without examples or specific training on that type of question.