Reference

AI Visibility Glossary

The vocabulary that comes up in audits, reports, and the help center — defined in plain language.

24 terms

AEO (Answer Engine Optimization)

Optimising content so AI assistants can quote it when answering a question, as opposed to optimising for a ranked list of links.

AI Visibility Score

A 0–100 summary of how legible a site is to AI answer engines, built from 32 binary checks across four weighted pillars. Read more →

Answer engine

A system that responds with a synthesised answer rather than links — ChatGPT, Claude, Gemini, and Perplexity being the common examples.

Answer-first writing

Leading a section with the direct answer, then supporting detail. Makes a passage self-contained and therefore quotable.

Authority Signals

The pillar covering proof: testimonials, case studies, founder presence, review markup, and external corroboration. Worth 20% of the score. Read more →

Answer Selection

The pillar covering whether your content is shaped the way answers get assembled — FAQ content, comparisons, question-shaped headings. Worth 30%.

ChatGPT-User

OpenAI's user-triggered fetcher. Visits a page when a user asks about it, distinct from the training crawler GPTBot. Read more →

Citation

When an assistant attributes part of its answer to a specific source page. Happens at passage level, not site level.

ClaudeBot

Anthropic's crawler associated with collecting training data. Controlled separately from user-triggered fetches.

Entity confidence

Whether an AI system can tell your brand apart from similar names and state what it is without hedging. Read more →

Evidence Ladder

Surfaced's five-tier grading of proof, from unverified claims at the bottom to original data at the top. Read more →

FAQPage schema

Schema.org markup that states question–answer pairs explicitly, making them straightforward for a system to lift.

GPTBot

OpenAI's training-data crawler. Blocking it is a content decision that does not directly affect whether you are recommended today.

Google-Extended

A robots.txt token controlling use of your content for Google's generative AI training, separate from Search indexing.

Grounding

Anchoring a generated answer in retrieved sources rather than model memory alone.

llms.txt

A proposed plain-text file at a site root that summarises what the site is and points to key pages. Read more →

OAI-SearchBot

OpenAI's indexing crawler for ChatGPT search. Blocking it removes you from that index.

Passage-level retrieval

Retrieving and citing a specific chunk of a page rather than the whole document. Why self-contained sections matter.

PerplexityBot

Perplexity's indexing crawler, distinct from Perplexity-User which fetches pages on demand.

RAG (Retrieval-Augmented Generation)

Fetching relevant documents at answer time and generating a response grounded in them.

Schema.org

A shared structured-data vocabulary. Used to state facts about a page explicitly instead of leaving them to inference.

Structural Understanding

The pillar covering whether your copy tells a machine what you are and who you serve. Worth 30% — the joint-largest.

Technical Visibility

The pillar covering whether crawlers can reach and parse your site at all: HTTPS, robots.txt, sitemap, schema, llms.txt. Worth 20%.

User-triggered fetcher

A crawler that only visits a page because a user asked about it. Blocking these removes you from live answers. Read more →