GEO and AI search glossary — terms, explained

What this category covers

How a generative system gets from a question to an answer with sources attached — and which parts of that path a publisher can influence. Newer than the SEO vocabulary, and less settled: where a term is contested, the page says so.

Retrieval — how a passage is found before it is quoted
Chunking and retrieval fit — why a page is read in pieces
Grounding, citations and hallucination
Prompts, query fan-out and how one question becomes several
Citation share and visibility in answers
The difference between traffic and being quoted

Why these words matter

The answer is the surface now

A reader who gets a complete answer may never see a list of links. Whether your page is in that answer is a different question from where it ranks, and it needs its own words.

It is measurable, mostly not by the old tools

Impressions and positions say nothing about a citation. The terms here name what can actually be counted: how often you are quoted, beside whom, for which questions.

The mistakes are new

Blocking the crawler that feeds an assistant, or writing for a model instead of a reader, are errors that did not exist five years ago and are easy to make by accident.

Who needs this vocabulary

SaaS

Buyers ask an assistant to compare tools, and the comparison is assembled from whatever it can retrieve.

E-commerce

Product questions are answered before the shop is opened, from pages the shop did not write.

Regulated industries

Finance, health, legal — where a confidently wrong answer about you is a real cost.

B2B services

Few queries, high value, and a shortlist that an assistant is now often the one to draw up.

Publishers

Content is the product, and being summarised without attribution is the failure mode.

Agencies

Clients are asking about AI visibility already; the words have to mean something before they are billed for.

Terms in this category319 terms

A40

Accountability5 min read Action Space5 min read ads.txt7 min read Adversarial Attack4 min read AEO6 min read AEO5 min read Agent Loop7 min read Agent Orchestration5 min read Agent Persona6 min read Agent Protocol6 min read Agent Readiness5 min read Agent Simulator5 min read Agent Skill6 min read Agentic Browsing3 min read Agentic Commerce5 min read Agentic Search5 min read AI Assistant4 min read AI Audit5 min read AI Cloaking5 min read AI Ethics4 min read AI Governance4 min read AI Mode4 min read AI Overview3 min read AI Overviews5 min read AI Visibility4 min read Amazon Alexa6 min read Annoy2 min read Answer Engine6 min read Anthropic API6 min read API Key6 min read app-ads.txt6 min read Apple Siri4 min read Assistant6 min read Assistant Surface7 min read Attention Mechanism5 min read Autogen5 min read Autonomous Agent6 min read AWS Bedrock4 min read AXE5 min read Azure AI5 min read

B9

Beam Search7 min read BERT4 min read Bi-Encoder6 min read Bias Mitigation6 min read Bing Chat4 min read BLEU6 min read BM253 min read Brandscope7 min read Brave Search5 min read

C28

D13

Data Augmentation7 min read Data Cleaning5 min read Data Poisoning5 min read Data Preprocessing5 min read Data Sheet for Datasets5 min read DeepSeek5 min read Dense Retrieval5 min read Deployment7 min read Differential Privacy5 min read Diversity Ranking4 min read Document Store5 min read DocX7 min read Domain Adaptation5 min read

E6

Elasticsearch7 min read Embedding6 min read Encoder-Decoder5 min read Entity5 min read Entity Extraction4 min read Entity Salience5 min read

F13

F1 Score6 min read Factuality5 min read Fairness6 min read Faiss5 min read FCP5 min read Feature Engineering5 min read Feature Importance5 min read Federated Learning7 min read Feedback Loop4 min read Few-shot Learning6 min read Few-shot Prompting6 min read Fine-tuning5 min read Function Calling7 min read

G15

Gemma4 min read Generative Search Engine5 min read GEO5 min read GEO4 min read GLM5 min read Golden Dataset4 min read Google AI Studio5 min read Google Antigravity5 min read Google Assistant6 min read Google Gemini6 min read GPT4 min read Grok5 min read Grounding5 min read GSE5 min read Guardrails5 min read

H11

Hacker News5 min read Hallucination in AI Search5 min read Hermes5 min read Hnswlib6 min read HTML3 min read HTML Meta Tags7 min read Hugging Face Hub4 min read Hugging Face Transformers5 min read Human Evaluation6 min read Hybrid Retrieval5 min read Hybrid Search4 min read

I6

In-memory Database6 min read Inference7 min read Inference Endpoint3 min read Information Gain5 min read Intent Recognition7 min read Interpretability5 min read

J4

Jailbreak4 min read Jan.ai5 min read JSON5 min read JSON-LD7 min read

K6

Kagi5 min read Kimi4 min read Knowledge Cutoff4 min read Knowledge Distillation5 min read Knowledge Graph6 min read Knowledge Retrieval4 min read

L20

M23

N8

NDCG5 min read Nemotron5 min read NLG5 min read NLP5 min read NLU5 min read Normalization3 min read NotebookLM4 min read Nucleus Sampling7 min read

O10

Observation Space7 min read Obsidian5 min read Ollama5 min read Open Graph Protocol6 min read Open Interpreter7 min read OpenAI API7 min read OpenAI Embeddings7 min read OpenCode5 min read OpenRouter4 min read OpenSearch5 min read

P21

Q6

QAS5 min read Qdrant4 min read Quantization6 min read Query Expansion4 min read Query Fan-Out4 min read Qwen4 min read

R19

R@K6 min read RAG5 min read Rate Limiting6 min read Re-ranking Model7 min read Reasoning Model6 min read Reddit5 min read Reflexion6 min read Reinforcement Learning6 min read Rel=Canonical7 min read Reranking7 min read Responsible AI5 min read Retrieval4 min read Retrieval Pipeline5 min read RLAIF5 min read RLHF7 min read Robustness5 min read ROUGE7 min read RRF7 min read RSL5 min read

S26

T16

Test-Time Compute6 min read Text-to-Speech5 min read TF-IDF6 min read Token6 min read Tokenization4 min read Tool Integration6 min read Tool Registry7 min read Tool Use5 min read Top-k Sampling4 min read ToT4 min read Transfer Learning5 min read Transformer Model4 min read Transparency Report5 min read Truthfulness4 min read Twitter Cards6 min read TXT4 min read

U1

User Prompt3 min read

V8

Value Alignment5 min read Vector Database7 min read Vector Embedding5 min read Vector Search4 min read Vector Store6 min read Vectorscope7 min read Vision-Language Model5 min read Voicescope7 min read

W2

Weaviate5 min read web-manifest6 min read

X4

XAI5 min read XLSX4 min read XML Sitemap5 min read XRL5 min read

Y1

You.com5 min read

Z3

Zero-Click Search6 min read Zero-shot Learning7 min read Zero-shot Prompting4 min read

How to use this category

1

Learn the shape of an answer

Start with retrieval, chunking and citation. Everything else on this page assumes those three.

2

Find out where you stand

A term becomes useful the moment you ask it of your own brand: are we quoted for this question, and who is quoted instead?

3

Fix the page that should have been cited

Usually it exists and is unquotable — buried lead, no structure, no source. The terms here name each of those faults.

Common mistakes

Assuming a first-place ranking means a citation. They are measured separately because they behave separately.
Writing for a model instead of a reader — which produces pages that read as filler to both.
Blocking the crawlers that feed the assistants, then wondering why nothing cites you.
Counting mentions without reading them. Being quoted as a bad example is a mention.

Questions about this category

What does GEO actually stand for?

Generative engine optimisation — the work of being retrievable and quotable by systems that answer in prose. The term is a few years old and not universal; some people say AEO, answer engine optimisation, for the same thing.

Does GEO replace SEO?

No. It sits on top of it. The retrieval an assistant does is only possible on pages a crawler can already reach, so the older vocabulary stays load-bearing.

Can I make an assistant cite me?

No, and anybody who says otherwise is selling something. You can make a page easier to retrieve, easier to quote and harder to get wrong. What is cited is decided by a system you do not control.

How is visibility in AI answers measured?

By asking the assistants the questions your buyers ask, repeatedly, and recording what comes back — who was named, who was cited, and in what terms. That is what this product does, and the measurement vocabulary is its own category.

The other categories

Search engine optimisationThe words that describe how a page is found in a search engine — crawling, indexing, intent, structure, links. This is the ground the AI-search vocabulary is built on, not a separate subject.Trust and E-E-A-TExperience, expertise, authoritativeness and trust — the signals that decide whether a page is treated as a source or as noise, by a search engine and by a system writing an answer.MeasurementThe words for counting what happened — impressions, share of voice, coverage, sampling, windows, confidence — and, for each of them, what it cannot tell you.Marketing and growthThe commercial vocabulary around search — what an auction actually reports, who your real competitors are, how attention turns into revenue, and the frameworks used to decide what to write next.

See whether assistants quote you

Reading a definition tells you what the word means. The Vectorscope tells you what your own site looks like to the systems that use it.

Open the Vectorscope