aeo-geo-optimizer

AEO/GEO Optimizer

Optimize content and websites for AI-powered search experiences — ChatGPT, Perplexity, Claude, Google AI Overviews, and Bing Copilot.

Install

git clone https://github.com/thatrebeccarae/claude-marketing.git && cp -r claude-marketing/skills/aeo-geo-optimizer ~/.claude/skills/

Why This Matters

Traditional SEO optimizes for 10 blue links. AEO/GEO optimizes for AI-generated answers. When someone asks ChatGPT or Perplexity a question, the answer synthesizes from sources — and those sources get cited, linked, and trusted. If your content is not structured for AI consumption, you are invisible in the fastest-growing search channel.

Core Concepts

AEO vs GEO vs Traditional SEO

| Dimension | Traditional SEO | AEO (Answer Engine) | GEO (Generative Engine) | |-----------|----------------|---------------------|------------------------| | Target | Google/Bing SERPs | Featured snippets, AI Overviews, voice assistants | ChatGPT, Perplexity, Claude citations | | Goal | Rank on page 1 | Be THE answer | Be cited in AI-generated responses | | Content format | Long-form, keyword-rich | Concise, structured Q&A | Authoritative, quotable, fact-dense | | Signals | Backlinks, keywords, UX | Schema markup, direct answers, authority | E-E-A-T, data density, citation-worthiness | | Measurement | Rankings, traffic | Answer box appearance, voice search hits | AI citation tracking, brand mentions in AI |

The Citation Hierarchy

AI models prioritize sources based on:

1. Authority signals — Domain authority, author expertise, institutional backing
2. Content structure — Clear headings, direct answers, structured data
3. Freshness — Recent publication dates, updated statistics
4. Specificity — Exact numbers, named sources, verifiable claims
5. Uniqueness — Original research, proprietary data, novel frameworks

AEO/GEO Audit Workflow

Step 1: Assess Current AI Visibility

1. Test AI citation presence: Query ChatGPT, Perplexity, and Google AI Overviews with questions your content should answer. Document which queries cite your content vs competitors.

1. Check structured data: Validate schema markup coverage using Google Rich Results Test or Schema.org validator.

1. Evaluate content structure: Score each page on AEO readiness using the Content Scorecard below.

Step 2: Content Scorecard

Rate each piece of content (1-5) on these dimensions:

| Dimension | Score 1 (Poor) | Score 5 (Excellent) | |-----------|---------------|-------------------| | Direct answers | Buried in paragraphs | Clear Q&A format, first-sentence answers | | Data density | Opinions without evidence | Specific numbers, percentages, dates | | Source attribution | No citations | Named sources, linked studies | | Structure | Wall of text | H2/H3 hierarchy, lists, tables | | Schema markup | None | Article, FAQ, HowTo, or relevant type | | Freshness signals | No dates | Published date, "Updated" date, recent data | | Author authority | No byline | Named author with expertise credentials | | Quotability | Meandering prose | Crisp, self-contained statements AI can extract |

Scoring: 32-40 = AI-ready. 24-31 = Needs optimization. Below 24 = Major rework needed.

Step 3: Optimize for AI Citation

Content Structure Patterns

The Direct Answer Pattern:

## [Question as H2]

[One-sentence direct answer.] [Supporting context in 2-3 sentences.]

**Key details:**
- [Specific data point]
- [Specific data point]
- [Source attribution]

The Definition Pattern:

## What Is [Term]?

[Term] is [clear, concise definition in one sentence]. [Elaboration with context.] [How it differs from related concepts.]

The Comparison Pattern:

## [X] vs [Y]: Key Differences

| Dimension | [X] | [Y] |
|-----------|-----|-----|
| [Aspect 1] | [Specific detail] | [Specific detail] |
| [Aspect 2] | [Specific detail] | [Specific detail] |

**Bottom line:** [One-sentence recommendation with reasoning.]

The Statistics Pattern:

## [Topic] Statistics ([Year])

- **[Stat 1]**: [Number] ([Source, Year])
- **[Stat 2]**: [Number] ([Source, Year])
- **[Stat 3]**: [Number] ([Source, Year])

*Sources: [List with links]*

Writing for AI Extraction

1. Lead with the answer. AI models extract the first sentence after a heading. Make it count.
2. Use specific numbers. "Revenue increased 47% year-over-year" beats "revenue increased significantly."
3. Name your sources. "According to a 2026 McKinsey report" is citable; unsourced claims are not.
4. Create self-contained paragraphs. Each paragraph should make sense extracted in isolation.
5. Use comparison tables. AI models love structured comparisons — they are easy to synthesize.
6. Include "What is" and "How to" headings. These directly match common AI queries.
7. Add freshness signals. Include publication date, last-updated date, and date-stamp your statistics.
8. Write quotable sentences. Crisp, declarative statements that AI can extract verbatim.

Technical Optimization

1. Schema markup — See the schema-markup-generator skill for implementation
2. Canonical URLs — Ensure AI models find the authoritative version
3. XML sitemap — Keep it current so AI crawlers find new content
4. Page speed — AI crawlers respect crawl budgets; fast sites get crawled more
5. robots.txt — Ensure AI crawlers (GPTBot, anthropic-ai, PerplexityBot) are not blocked

Step 4: Monitor AI Visibility

AI Crawler User Agents

| Crawler | User Agent | Purpose | |---------|-----------|---------| | OpenAI | GPTBot | ChatGPT training and browsing | | Anthropic | anthropic-ai, ClaudeBot | Claude training and citations | | Perplexity | PerplexityBot | Perplexity search citations | | Google | Google-Extended | Gemini/AI Overview training | | Microsoft | Bingbot (+ AI signals) | Bing Copilot citations | | Meta | Meta-ExternalAgent | Meta AI features | | Apple | Applebot-Extended | Apple Intelligence |

robots.txt Recommendations

# Allow AI crawlers for maximum AI search visibility
User-agent: GPTBot
Allow: /

User-agent: anthropic-ai
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Google-Extended
Allow: /

Decision framework: If your goal is AI visibility (AEO/GEO), allow all AI crawlers. If you have licensing concerns about training data, selectively block training-only crawlers while allowing search/citation crawlers.

Measurement Approaches

| Method | What It Tracks | Tools | |--------|---------------|-------| | Manual citation checks | Query AI platforms, document citations | ChatGPT, Perplexity, Google | | Server log analysis | AI crawler frequency and pages crawled | Log analyzers, custom scripts | | Brand mention monitoring | Your brand/content mentioned in AI answers | Manual checks, brand monitoring tools | | Referral traffic | Traffic from AI platforms | GA4 (check referral sources for chat.openai.com, perplexity.ai) | | Schema validation | Structured data coverage and errors | Google Search Console, Rich Results Test |

Content Types and AI Optimization

Blog Posts / Articles

• Add FAQ schema for common questions
• Structure with clear H2 question headings
• Include "Key Takeaways" or "TL;DR" section
• Date-stamp all statistics

Product / Service Pages

• Add Product or Service schema
• Include comparison tables vs alternatives
• Answer "What is [product]?" in first paragraph
• List specific features with quantified benefits

Documentation / How-To Content

• Add HowTo schema with explicit steps
• Number every step
• Include time estimates and difficulty level
• Add "Prerequisites" and "Common Mistakes" sections

Research / Data Content

• Add Dataset schema where applicable
• Lead with key findings before methodology
• Create a "Key Statistics" summary section
• Ci