AEO GEO optimizer
Quick answer
- 01What is it?
- Optimize content and websites for AI-powered search experiences, ChatGPT, Perplexity, Claude, Google AI Overviews, and Bing Copilot. The value is a focused slice of search and SEO workflows judgment, useful when several similar skills cover the same ground.
- 02Inputs
- Context for search and SEO workflows: your goals, audience, constraints, and any source material the skill asks for.
- 03Output
- A ready-to-use result for search and SEO workflows: the analysis, copy, or recommendations the agent produces.
Add this skill
Install as a package
Installs this one skill package for your coding agent, including any supporting files that skill ships with — not every skill in the repository. Read the tutorial.
$ npx skills add thatrebeccarae/claude-marketing --skill aeo-geo-optimizerSkill instructions
The instruction file for this skill. The skill also includes other files you need to install to use it.
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:
- Authority signals — Domain authority, author expertise, institutional backing
- Content structure — Clear headings, direct answers, structured data
- Freshness — Recent publication dates, updated statistics
- Specificity — Exact numbers, named sources, verifiable claims
- Uniqueness — Original research, proprietary data, novel frameworks
AEO/GEO Audit Workflow
Step 1: Assess Current AI Visibility
-
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.
-
Check structured data: Validate schema markup coverage using Google Rich Results Test or Schema.org validator.
-
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
- Lead with the answer. AI models extract the first sentence after a heading. Make it count.
- Use specific numbers. "Revenue increased 47% year-over-year" beats "revenue increased significantly."
- Name your sources. "According to a 2026 McKinsey report" is citable; unsourced claims are not.
- Create self-contained paragraphs. Each paragraph should make sense extracted in isolation.
- Use comparison tables. AI models love structured comparisons — they are easy to synthesize.
- Include "What is" and "How to" headings. These directly match common AI queries.
- Add freshness signals. Include publication date, last-updated date, and date-stamp your statistics.
- Write quotable sentences. Crisp, declarative statements that AI can extract verbatim.
Technical Optimization
- Schema markup — See the schema-markup-generator skill for implementation
- Canonical URLs — Ensure AI models find the authoritative version
- XML sitemap — Keep it current so AI crawlers find new content
- Page speed — AI crawlers respect crawl budgets; fast sites get crawled more
- 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-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
- Cite sample sizes, date ranges, and confidence levels
Anti-Patterns (Never Do)
- Do not block AI crawlers if your goal is AI visibility
- Do not write "click here" or "read more below" — AI extracts content out of context
- Do not bury answers in long introductions — lead with the answer
- Do not use vague qualifiers — "many," "significant," "some" — use specific numbers
- Do not neglect author bylines — E-E-A-T signals matter to AI models
- Do not duplicate content across pages — AI models deduplicate and may ignore both
- Do not over-optimize for one AI platform — optimize for all of them
- Do not forget internal linking — AI crawlers follow links to build topical authority maps
Integration with Other Skills
- technical-seo-audit — Run technical audit first, then layer AEO/GEO optimization
- schema-markup-generator — Generate the structured data this skill recommends
- seo-content-writer — Apply AEO writing patterns during content creation
- content-creator — Use brand voice analysis to maintain voice while optimizing for AI
Supporting file: EXAMPLES.md
AEO/GEO Optimizer — Examples
Example 1: Full AEO/GEO Audit
Prompt
Audit our marketing blog for AI search visibility. We publish 3 articles per week about email marketing and want to appear in ChatGPT and Perplexity answers.
What the skill does
-
Assess current visibility: Tests 10 target queries across ChatGPT, Perplexity, and Google AI Overviews. Documents which queries cite the blog vs competitors.
-
Crawlability check: Reviews robots.txt for AI crawler access. Finds GPTBot and PerplexityBot are blocked — flags as critical issue.
-
Content scorecard: Scores the 5 most recent articles using the AEO scoring rubric:
- Article 1: 18/40 — no dates, no author byline, no schema
- Article 2: 24/40 — has dates but answers buried in paragraph 3
- Article 3: 21/40 — good structure but no statistics cited
-
Produces prioritized recommendations:
- Critical: Unblock AI crawlers in robots.txt
- High: Add Article + Person schema to all blog posts
- High: Restructure content to lead with direct answers
- Medium: Add FAQ sections with related questions
- Medium: Date-stamp all statistics with sources
Example 2: Optimize Existing Article for AI Citation
Prompt
Optimize this article about email deliverability best practices so it gets cited by AI search engines.
What the skill does
-
Scores current state: Runs the Content Scorecard — scores 22/40
-
Identifies specific issues:
- No direct answer in first paragraph (starts with background)
- Statistics without sources ("studies show" instead of named studies)
- No FAQ schema markup
- H2 headings are topic-based, not question-based
-
Rewrites key sections following AEO patterns:
- Adds "What is email deliverability?" H2 with one-sentence definition
- Restructures "Best Practices" into numbered steps for HowTo schema
- Adds comparison table: Good vs Poor deliverability signals
- Creates FAQ section with 5 questions
- Adds specific statistics with sources and dates
-
Generates schema markup (references schema-markup-generator skill)
Example 3: Competitive AI Visibility Analysis
Prompt
Our competitor keeps getting cited in ChatGPT for email marketing questions. What are they doing that we are not?
What the skill does
- Queries 15 email marketing questions in ChatGPT and Perplexity
- Documents which sources are cited for each query
- Analyzes competitor content structure vs yours:
- Competitor uses question-format H2s — you use topic-format
- Competitor includes specific benchmarks with sources — you use general statements
- Competitor has FAQ schema on every page — you have none
- Competitor content is updated quarterly with fresh dates — yours has no update dates
- Produces a gap analysis with specific remediation steps for each finding
Supporting file: REFERENCE.md
AEO/GEO Optimizer — Reference
AI Platform Citation Behaviors
ChatGPT (OpenAI)
- Browsing mode: Cites sources with links when browsing is enabled
- Training cutoff: Knowledge up to training date; browsing supplements
- Citation style: Inline links with source titles
- Preferences: Authoritative domains, recent content, structured data
- Crawler: GPTBot (respect robots.txt)
Perplexity
- Always cites: Every response includes numbered source citations
- Real-time search: Fetches current web results for every query
- Citation style: Numbered footnotes with full URLs
- Preferences: Content that directly answers the query, recent publications
- Crawler: PerplexityBot
Google AI Overviews
- Integrated into SERPs: Appears above organic results
- Citation style: Source cards with site name and favicon
- Preferences: E-E-A-T signals, schema markup, topical authority
- Trigger: Informational and "how-to" queries
- Crawler: Googlebot + Google-Extended for AI features
Claude (Anthropic)
- Tool use: Can browse web when tools are available
- Citation style: Inline references to source content
- Preferences: Well-structured, authoritative, factual content
- Crawler: anthropic-ai, ClaudeBot
Schema Markup for AI Visibility
Priority Schema Types for AEO
| Content Type | Schema | AI Benefit |
|---|---|---|
| Articles | Article, NewsArticle | Author credibility, publish date |
| How-tos | HowTo | Step extraction for procedural queries |
| FAQs | FAQPage | Direct Q&A matching |
| Products | Product | Feature/price comparison queries |
| Reviews | Review, AggregateRating | Trust signals |
| Data/Stats | Dataset | Factual citation preference |
| People | Person | Author expertise signals |
| Organizations | Organization | Brand authority |
Recommended Schema Combinations
Blog Post: Article + Person (author) + BreadcrumbList + FAQPage (if FAQ section)
Product Page: Product + AggregateRating + FAQPage + BreadcrumbList
How-To Guide: HowTo + Article + Person + BreadcrumbList
Content Optimization Checklists
Article Optimization for AI Citation
- H1 contains the primary question/topic
- First paragraph directly answers the core question
- All statistics include source and year
- Author byline with credentials/expertise
- Published date and last-updated date visible
- FAQ section with 3-5 related questions
- Comparison table (if applicable)
- Key Takeaways / TL;DR section
- Article schema markup implemented
- Internal links to related authoritative content
Page-Level AEO Checklist
- Title tag includes target question/topic (50-60 chars)
- Meta description answers the query concisely (150-160 chars)
- H2 headings use question format where natural
- Content is scannable (short paragraphs, lists, tables)
- Images have descriptive alt text
- Page loads in under 2.5 seconds (LCP)
- Mobile-responsive layout
- Canonical URL set correctly
- AI crawlers allowed in robots.txt
Query Intent Mapping
| Intent Type | AI Behavior | Content Strategy |
|---|---|---|
| Definitional ("What is X?") | Extracts first clear definition | Lead H2 with definition, keep under 2 sentences |
| Procedural ("How to X") | Extracts numbered steps | Use HowTo schema, number every step |
| Comparative ("X vs Y") | Synthesizes from comparison content | Comparison tables, explicit pros/cons |
| Statistical ("How many X?") | Extracts specific numbers | Bold key stats, cite sources inline |
| Evaluative ("Best X for Y") | Aggregates recommendations | Clear criteria, ranked recommendations |
| Explanatory ("Why does X?") | Extracts causal reasoning | Direct cause-effect in first paragraph |
Measurement Framework
Tracking AI Citations
Manual Process (Monthly):
- Compile your top 20 target queries
- Ask each AI platform (ChatGPT, Perplexity, Google AI Overview)
- Record: Was your content cited? Which competitor was cited instead?
- Track month-over-month citation rate
Automated Signals:
- GA4 referral traffic from:
chat.openai.com,perplexity.ai,you.com - Server logs: crawl frequency from GPTBot, PerplexityBot, ClaudeBot
- Search Console: AI Overview impressions (where available)
- Brand monitoring: mentions in AI-generated content
KPIs
| Metric | How to Measure | Target |
|---|---|---|
| AI citation rate | Manual query testing | >30% of target queries |
| AI referral traffic | GA4 referral report | Month-over-month growth |
| Crawl frequency | Server log analysis | Weekly crawls from AI bots |
| Schema coverage | Rich Results Test | 100% of eligible pages |
| Content score | AEO Scorecard (above) | >32/40 on key pages |
Common questions
How do I install AEO GEO optimizer in Cursor, Claude Code, or Codex?
Run npx skills add thatrebeccarae/claude-marketing --skill aeo-geo-optimizer in the project where you want it, then ask your agent for the skill by name. The --skill flag installs only AEO GEO optimizer, not every skill in the repository.
Where does AEO GEO optimizer come from and what license is it under?
AEO GEO optimizer comes from the thatrebeccarae/claude-marketing repository on GitHub. That repository has 67 GitHub stars. The skill is published under the MIT license.
Prefer plain text? Read the AEO GEO optimizer guide as markdown.
Related skills
More from thatrebeccarae