The vocabulary, defined once, clearly.

A

AI crawler

AI bot, LLM crawler

An AI crawler is an automated agent that fetches web pages on behalf of an AI system and identifies itself with its user-agent string, such as GPTBot, ClaudeBot or PerplexityBot. Vendors separate crawlers by purpose — collecting training data, building a search index, or fetching a page a user asked about — so that training and search access can be allowed or blocked independently in robots.txt.Full entry

AI Overviews

Google AI Overviews, formerly Search Generative Experience (SGE)

AI Overviews is a Google Search feature that shows an AI-generated summary at the top of some results pages, with links to the web pages the summary drew on. It launched to all US users in May 2024 after a year as the Search Generative Experience (SGE) experiment in Search Labs, and is fed by the same crawling and indexing as ordinary Search.Full entry

Answer engine

An answer engine is a system that responds to a question with a synthesised answer rather than a ranked list of links. Modern answer engines — Google AI Overviews and AI Mode, Perplexity, ChatGPT search, Claude with web search and Microsoft Copilot — retrieve passages from the web and generate the answer with a large language model, usually with citations to the sources used.Full entry

Answer Engine Optimization

AEO

Answer Engine Optimization (AEO) is the practice of structuring web content so that systems which return a direct answer — AI assistants, AI-powered search and voice assistants — can retrieve, understand and cite it. It prioritises self-contained answer passages, clear entities, verifiable authorship and machine-readable formats over ranking position.Full entry

C

Canonical URL

rel=canonical, canonical link

A canonical URL is the single address a site declares as the authoritative version of a page when the same content is reachable at several URLs — with and without a trailing slash, with tracking parameters, as a Markdown or syndicated copy. It is declared with a link element whose rel attribute is canonical, or the equivalent HTTP Link header, and tells crawlers where to consolidate signals and attribution.Full entry

Chunking

Text splitting, passage segmentation

Chunking is the step in a retrieval pipeline that splits a page or document into smaller passages — typically a few hundred tokens each, cut at headings, paragraphs or a fixed window — so that each passage can be embedded, indexed and retrieved on its own. Because answer engines retrieve and cite chunks rather than whole pages, chunk boundaries decide what a model actually sees.Full entry

Citation

Source attribution, reference

A citation, in the context of answer engines, is the explicit link or reference that a generated answer attaches to the source passage it drew on — the numbered footnote in Perplexity, the linked sources beside a Google AI Overview, or the source list under a ChatGPT search answer. Citations are the unit of visibility that Answer Engine Optimization is measured by.Full entry

Content negotiation

Proactive negotiation, server-driven negotiation

Content negotiation is the HTTP mechanism by which a client states which representations it prefers — using headers such as Accept, Accept-Language and Accept-Encoding — and the server picks the best available one for the same URL. For answer engines it lets one canonical URL return HTML to browsers and Markdown to a crawler that sends Accept: text/markdown.Full entry

D

DefinedTerm

schema.org/DefinedTerm, DefinedTermSet

DefinedTerm is a Schema.org type for a word, name, acronym or phrase that has a formal definition. It carries a name, a description that holds the definition, an optional termCode, and an inDefinedTermSet link to the DefinedTermSet — the glossary or classification — it belongs to. It is the natural markup for a glossary entry and tells a machine that a passage is a definition.Full entry

E

E-E-A-T

Experience, Expertise, Authoritativeness, Trustworthiness; formerly E-A-T

E-E-A-T stands for Experience, Expertise, Authoritativeness and Trustworthiness: the framework Google's Search Quality Rater Guidelines use to describe whether a page and its author deserve to be believed. It is not a direct ranking signal but a description of the qualities Google's ranking systems try to reward, with trust named as the most important of the four.Full entry

Embedding

Vector embedding, text embedding

An embedding is a fixed-length list of numbers (a vector) that a neural model produces to represent the meaning of a piece of text, so that texts with similar meaning end up close together in the vector space. Answer engines embed both the question and candidate passages, then retrieve the passages whose vectors are nearest to the question's.Full entry

Entity

An entity is a uniquely identifiable thing — a person, organisation, product, place or concept — that exists independently of the words used to name it. Search engines and language models resolve text to entities so that 'Apple', 'Apple Inc.' and 'AAPL' map to one node, and they judge a page partly by which entities it mentions and how unambiguously it names them.Full entry

G

Generative Engine Optimization

GEO

Generative Engine Optimization (GEO) is the practice of increasing how often, and how favourably, a source is used and cited by generative engines — search systems that answer with text produced by a large language model, such as Google AI Overviews, Perplexity or ChatGPT search. The term was introduced in a 2023 research paper by Aggarwal et al.Full entry

Grounding

Grounding is the practice of tying a language model's output to specific, verifiable sources — retrieved web pages, documents or a knowledge graph — supplied at generation time, so that each claim in the answer can be traced to evidence rather than to the model's training data. In answer engines, grounding is what turns a retrieved passage into a cited one.Full entry

H

Hallucination

Confabulation, fabrication

A hallucination is output from a language model that is fluent and confident but false or unsupported: an invented statistic, a citation to a paper that does not exist, a misattributed quote, or a plausible-sounding date. It arises because the model predicts likely text rather than checking facts, and it is the main failure that answer engines use retrieval and citations to contain.Full entry

Headless CMS

API-first CMS, decoupled content platform

A headless CMS is a content management system that stores content as structured fields and delivers it over an API, leaving the presentation layer — website, app, feed, Markdown export — to be built separately. For answer engines it matters because the same fields that render the page can also generate JSON-LD, Markdown and llms.txt without the outputs drifting apart.Full entry

J

JSON-LD

JSON for Linking Data, application/ld+json

JSON-LD (JSON for Linking Data) is a W3C standard for expressing linked data as ordinary JSON. On the web it is the format Google recommends for Schema.org structured data: a single script block of type application/ld+json that describes the page's entities — article, author, dates, FAQ, definitions — without touching the visible HTML.Full entry

K

Knowledge graph

A knowledge graph is a structured database of entities — people, organisations, places, products, concepts — and the typed relationships between them, such as founded by, located in or is a subtype of. Search engines and answer engines use one to disambiguate names, verify facts and assemble panels; Google's Knowledge Graph, introduced in 2012, and the open Wikidata project are the best-known examples.Full entry

L

llms.txt

llms.txt is a proposed convention for a Markdown file at the root of a website (/llms.txt) that gives large language models a curated, plain-text overview of the site: a title, a short summary, and sections of links to the most important pages with one-line descriptions. A companion file, llms-full.txt, contains the full text of those pages in one document.Full entry

P

Passage retrieval

Passage retrieval is the search technique of scoring and returning short spans of text — a paragraph, a section, a window of tokens — rather than whole documents, so that a system can answer a question from the most relevant fragment. It is the retrieval step in most answer engines, and it is why a page is cited passage by passage rather than as a unit.Full entry

R

Retrieval-augmented generation

RAG

Retrieval-augmented generation (RAG) is an architecture in which a language model is given relevant documents or passages retrieved at query time — from a search index, a vector database or the live web — and asked to answer using them. It grounds the model's output in current, citable sources instead of relying only on what it memorised during training.Full entry

S

Schema.org

Schema.org is a shared vocabulary of types and properties — Article, Person, FAQPage, DefinedTerm and hundreds of others — for describing what a web page is about in a form machines can read. Founded in 2011 by Google, Microsoft, Yahoo and Yandex, it is maintained in the open through a W3C community group and can be written as JSON-LD, Microdata or RDFa.Full entry

Semantic HTML

Semantic HTML is markup that uses elements for what content is rather than how it should look: article, section, nav, h1–h6, dl, table, time, blockquote and cite, instead of generic div and span. Because crawlers and text extractors read structure rather than styling, semantic markup determines how a page is split into passages and which parts survive extraction.Full entry

Speakable

SpeakableSpecification

Speakable is a schema.org property (with the SpeakableSpecification type) that identifies, by CSS selector or XPath, which parts of a page are best suited to be read aloud or quoted verbatim — typically a headline and a short summary. It was introduced for voice assistants and remains the most direct way to tell a machine which passage is the answer.Full entry

Structured data

schema markup, semantic markup

Structured data, in the web context, is machine-readable markup embedded in a page that states explicitly what the page contains — an article, its author, a publication date, a question and its answer — using a shared vocabulary such as Schema.org. It is written as JSON-LD, Microdata or RDFa and lets crawlers and answer engines identify entities and relationships without inferring them from prose.Full entry

Z

Zero-click search

No-click search

A zero-click search is a query that ends on the results page without the searcher clicking through to any website, because the answer was shown directly — in a featured snippet, knowledge panel, AI Overview or a similar module. The term describes a measured outcome, not a feature: the searcher got what they needed, or gave up, and the source page received no visit.Full entry