# aeo

## What is AEO?

**AEO (Answer Engine Optimization)** is optimizing content for direct citation by "answer engines": Google AI Overviews, Yandex Neuro, ChatGPT, Perplexity. The core technique — put the ready answer in the first sentence and split the text into self-contained question-answer blocks a machine can quote without rewriting.

In the zero-click era, users often get the answer right in the results and never visit the site. AEO competes not for the click but for **source status**: the engine pulls your answer into its card and credits you with the citation.

## How is AEO different from SEO and GEO?

In short: SEO is about ranking pages, GEO is about a site's overall visibility in AI answers, and AEO is a specific tactic for delivering content in a ready-answer format. That's why AEO is usually treated as the practical part of [GEO](/glossary/geo).

| | Goal | Unit of optimization | Key technique |
|---|---|---|---|
| **SEO** | Rank in organic results | Page + keywords | Keywords, links, speed |
| **GEO** | Citation in AI answers | Whole site (content + technical base) | Structured data, llms.txt, robots AI-rules |
| **AEO** | Land in the "zero answer" | A single question-answer block | Inverted pyramid, Q&A, facts |

The mechanics differ: a search engine ranks a page by relevance, while an answer engine **extracts** a ready fragment from it. If the fragment can't be extracted, a competitor becomes the source.

## How do you write content for AEO?

Each section is built as an independent module that's easy for a machine to "slice." Six rules of the frame:

1. **Inverted pyramid.** The direct answer goes in the first sentence after the heading; details follow below.
2. **Heading = a user's question.** "How do I configure robots.txt for AI bots?" instead of "robots.txt setup."
3. **Short paragraphs.** One idea = up to 3–4 lines. Engines skip long walls of text.
4. **Lists and tables** instead of prose — structured data is parsed first.
5. **Concrete numbers.** "From 15 minutes," "10 of 23 checks" instead of "fast" and "many."
6. **Semantic cleanup.** Remove filler words — the meaning should survive without 70% of the adjectives.

A machine-readable layer amplifies the effect: mark up `FAQPage` or `HowTo` blocks in [Schema.org](/glossary/schema-org), serve clean [Markdown to agents](/glossary/markdown-for-agents), and describe key pages in [llms.txt](/glossary/llms-txt).

## How does AEO relate to our scanner?

There's no direct "AEO text quality" check in the scanner — that's a content metric that can't be measured reliably without judging meaning. But we do check the technical foundation of AEO directly:

- `schema_org` — whether structured markup exists (including `FAQPage`/`HowTo`)
- `markdown_negotiation` — whether clean Markdown is served to agents
- `llms_txt` — whether LLM navigation is published

These three checks show whether the page's infrastructure is ready for answer extraction. The text itself stays with the author — the six-rule frame above is the working checklist.

[← All glossary terms](/en/glossary)