The GEO Glossary - Common GEO Terminology Explained
WRITTEN BY
Kev Wiles
I’m a Fractional SEO Specialist with 12+ years’ experience working with eCommerce brands. I focus on making SEO simple, clear, and effective helping businesses cut through the noise and unlock real growth.
The GEO Glossary
There is a lot of chatter at the moment around Generative Engine Optimisation (GEO) but for those who are not close to the terminology or buzz words being thrown around I wanted to share a small overview of some of these terms and what they directly mean.
A lot of the GEO terminology I have seen used this year has been used pretty inconsistently. Some of it describes genuinely new mechanics in how AI systems find and cite content. Some of it is just SEO language with a new buzz word associated to it.
AEO (Answer Engine Optimisation) Structuring content so a specific passage can be lifted directly as the answer to a question, rather than merely ranked as a result. Often used interchangeably with GEO, but it's more specific: AEO is one of the techniques GEO relies on.
AI Mode Google's conversational search interface, separate from the standard results page. Instead of returning a page of links with a summary at the top, it functions as a full AI chat experience. The same inclusion factors that drive AI Overviews clear entities, strong authority signals, citable content apply here too.
AI Overviews Google's generated summaries that appear above the traditional results, built from multiple sources and citing them. Being one of the cited sources matters more here than holding a specific ranking position, because the overview itself is now the thing most people read.
Brand Entity Recognition The ability of AI systems to identify your brand as a distinct, verifiable, notable "thing" rather than just a string of text. This sits underneath a lot of AI visibility a brand an LLM can confidently identify and describe is far more likely to be recommended than one it can only match to a website.
Chunking The way an AI system breaks a page into smaller, retrievable units of text rather than treating the whole page as a single block. If a chunk doesn't stand on its own doesn't make sense pulled out of context.
Citation (AI) When a language model names your business or links your page as a source in its answer. Unlike a search ranking, this isn't a fixed position you hold the same question can produce different citations on different days, from different engines.
E-E-A-T Experience, Expertise, Authoritativeness and Trustworthiness. Google's framework for judging content quality. It isn't a score you can see directly; it's a description of the underlying signals named authors, verifiable claims, genuine first-hand experience, real credentials.
Entity A thing a search engine or language model recognises as distinct and can hold facts about a business, a person, a place, a product. Entity resolution is the process of a machine working out that different mentions ("Kev Wiles," "kevwiles.com," "Kevin Wiles SEO") all refer to the same thing.
Grounding When a language model bases its answer on retrieved documents rather than purely on what it learned during training. Grounded answers are the ones that cite sources which is exactly why having a crawlable, quotable page matters for GEO.
Hallucination When a language model states something false with confidence. For a brand, this most often shows up as an invented product spec, an out-of-date price, or a service you no longer offer which is why keeping authoritative source pages current is a GEO task, not just good housekeeping.
JSON-LD JavaScript Object Notation for Linked Data the format Google recommends for adding Schema.org markup to a page. It sits invisibly in the page's <head> and structures your data without touching how the page looks to a visitor.
Knowledge Graph A structured store of entities and the relationships between them. Google's version underpins the information panel that appears beside branded searches, and it's one of the sources LLMs draw on when they need to verify who or what a brand is.
LLM (Large Language Model) The class of model behind ChatGPT, Claude, Gemini, and similar systems. It generates text by predicting what comes next based on patterns in its training data, which is exactly why it can be fluent and wrong at the same time.
llms.txt A plain-text file at the root of a site giving AI systems a curated map of its most useful pages. It's a proposed convention, not an adopted standard no major AI provider has confirmed using it.
Passage-Level Optimisation Writing content in short, self-contained passages tight paragraphs, clear lists, well-labelled tables that an AI system can lift and cite without needing the surrounding context. The opposite of this is a long, meandering paragraph that only makes sense read in full from the top.
Prompt Volume How often a given question is actually asked of AI assistants. It isn't the same as search volume, and it's much harder to measure directly which is why a page can be worth writing even where Google Keyword Planner shows little to no demand.
Query Fan-out The process by which an AI system breaks one user prompt into several underlying sub-queries before retrieving information, then synthesises an answer from what each sub-query returns. A single question like "best ramp for a wheelchair user at home" might fan out into separate retrieval steps for slope regulations, product types, and installation guidance which means a page can be relevant to only part of a prompt's fan-out and still be pulled into the final answer.
RAG (Retrieval-Augmented Generation) A method where a model searches for relevant documents first, then writes its answer from what it finds. Most AI answer engines work this way, which means the retrieval step whether your page is found and judged relevant at all is what decides whether you get cited.
Robots.txt and AI Crawlers The file at your site's root that tells bots which pages they can and can't access increasingly relevant in GEO because it also governs the AI-specific crawlers. Common ones include GPTBot (OpenAI, for training future models), OAI-SearchBot (OpenAI, for real-time ChatGPT search citations a separate bot from GPTBot), ClaudeBot (Anthropic), PerplexityBot (Perplexity), and Google-Extended (controls use by Gemini and AI Overviews).
Schema / Structured Data Machine-readable markup added to a page describing what it actually contains a product, an article, a FAQ, a review.
Share of Voice (AI) / AI Visibility The proportion of relevant AI-generated answers in which your brand is mentioned, measured against competitors. It's the nearest equivalent to a rankings report for generative search and it's necessarily a sampled estimate, since the same prompt can return different answers on different days.
Kev Wiles © 2024 - Dubai Marina, Dubai,
