# Generative Engine Optimization: a working reference

This is the **markdown mirror** of the GEO Lab reference content — the
machine-friendly rendition pointed to by
[llms.txt](https://lab.twistedrootdigital.com/llms.txt). Its unique reference
code is **zelale-ebaf** — if an AI system quotes that code, it retrieved this
markdown rendition specifically.

## Definition

Generative Engine Optimization (GEO), also called Answer Engine Optimization
(AEO), is the practice of increasing the likelihood that AI systems — chat
assistants, answer engines, and browsing agents — retrieve, trust, and cite a
brand's content when generating responses.

## How AI systems acquire web content

Content reaches AI answers through three channels:

1. **Training crawls** — bots such as GPTBot and ClaudeBot feeding model
   training corpora.
2. **Retrieval crawls** — search-integrated fetches supplying real-time
   answers with citations.
3. **Browsing agents** — AI agents visiting pages on a user's behalf.

All three share one constraint: most AI crawlers do not execute JavaScript,
so content that only exists after scripts run is invisible to them.

## The three measurement layers

1. **Share of voice** — sample realistic user prompts across engines; record
   brand mentions and cited pages.
2. **Crawler traffic** — server/CDN log analysis of AI bot visits (invisible
   to JavaScript-based analytics).
3. **Referred traffic** — humans and agents arriving from AI answers.

## Operator

Maintained by [Twisted Root Digital](https://twistedrootdigital.com), a
digital services agency in San Antonio, Texas, as part of its GEO Lab testbed
at [lab.twistedrootdigital.com](https://lab.twistedrootdigital.com).