Core discipline terms

AEO (Answer Engine Optimization)
The practice of getting your brand mentioned and your content cited in AI-generated answers. Interchangeable with GEO. Full explainer.
GEO (Generative Engine Optimization)
Same discipline as AEO; the term comes from a 2023 academic paper. Emphasizes optimizing for generative models. Full explainer.
Answer engine
Any system that responds to a question with a synthesized answer rather than a list of links: ChatGPT, Claude, Perplexity, Google AI Overviews, Copilot.
AI visibility
The share of relevant AI-generated answers that mention your brand. The core AEO metric, usually tracked per prompt set, per engine, over time.
Share of voice
Your mentions relative to competitors’ across the same prompt set. More actionable than raw visibility because it controls for prompt difficulty.

How engines work

RAG (Retrieval-Augmented Generation)
The architecture behind most answer engines: retrieve relevant documents first, then generate an answer grounded in them. AEO is largely the art of winning the retrieval step.
Query fan-out
An engine decomposing one user question into multiple sub-queries before retrieving. "Best CRM for agencies" might fan out into pricing, review, and comparison sub-queries — each retrieving different pages.
Citation
A source link an engine attaches to its answer. Being cited means your page was retrieved and used; being mentioned without citation usually means the engine knows you from training data or third-party sources.
Training-data mention
A brand mention that comes from the model’s training corpus rather than live retrieval. Slow to change (updates with model releases) but durable once established.
Grounding
Tying generated claims to retrieved sources. Engines vary in how strictly they ground; stricter grounding means citations matter more.
Hallucination
A model stating something false with confidence — wrong pricing, features you don’t have, products that don’t exist. Monitoring for hallucinations about your brand is part of AEO hygiene.

Technical surface

AI crawler
A bot that fetches web content for AI systems: GPTBot and OAI-SearchBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, Google-Extended, CCBot (Common Crawl). Each has a distinct user agent you can allow or block in robots.txt.
llms.txt
A markdown index at your domain root that gives LLM agents a curated map of your site. Full guide.
llms-full.txt
Companion file to llms.txt carrying full page content for direct LLM consumption.
Content signals
robots.txt directives (search, ai-input, ai-train) declaring how your content may be used by search and AI systems.
Structured data / schema markup
JSON-LD annotations (Article, FAQPage, Organization) that make page meaning explicit to machines. Helps engines parse Q&A structure, dates, and authorship.
Markdown negotiation
Serving a markdown version of a page when a client sends Accept: text/markdown — token-efficient content delivery for agents. This site supports it.
MCP (Model Context Protocol)
An open protocol letting AI assistants call external tools and data sources. An MCP server makes your product’s data queryable by AI agents directly.
Agent-ready / agent-native
A site built so AI agents can discover and use it programmatically: crawler access, machine-readable indexes, discovery endpoints, and (at the high end) an MCP server. Cloudflare’s public scanner grades this.

Content patterns

Answer-shaped content
Pages structured for extraction: answer stated first, question-based headings, specific values, visible dates. The opposite of the 800-word wind-up intro.
Answer-first block
A self-sufficient 200-400 word passage at the top of a page that fully answers the target question — the chunk an engine can lift whole. (The highlighted box at the top of this page is one.)
Listicle
"Best X for Y" list content. Disproportionately cited for commercial queries because its format matches comparative questions. Parallel structure across entries matters.
E-E-A-T
Experience, Expertise, Authoritativeness, Trust — Google’s quality framework, relevant to AEO because credibility signals (real authors, citations, dates) influence which sources engines trust.
Citation decay
The tendency of a page’s AI citations to erode over time as fresher or better-matched sources enter retrieval. The reason AEO is monitoring, not a one-time project.