Audit and optimize content for AI search visibility across ChatGPT, Perplexity, Google AI Overviews, Claude, and Gemini. Analyze extractability, authority signals, and citation opportunities to maximize zero-click AI search presence.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
Audit and optimize content for AI search visibility across ChatGPT, Perplexity, Google AI Overviews, Claude, and Gemini. Analyze extractability, authority signals, and citation opportunities to maximize zero-click AI search presence.
role
Marketing
action
Analyze
name
ai-seo
connectors
[{"Google Drive":"Access existing blog posts, guides, and content docs to audit for AI extractability and structure"},{"Slack":"Tap team conversations to identify proprietary insights, customer language, and competitive positioning themes"},{"Gmail":"Mine customer conversations for authority angles, pain points, and language patterns AI models cite"},{"HubSpot":"Extract customer data and deal language to validate AI-optimized messaging aligns with buyer conversations"},{"Notion":"Reference content calendars and editorial guidelines to batch optimize upcoming content"}]
author
Harmony
version
1.0.0
icon
sparkles
cover
screenshots
[]
category
Marketing
Purpose
AI search engines (ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews) answer user queries by synthesizing and citing content from across the web. Unlike traditional SEO, AI Search Optimization (AEO) focuses on making your content extractable, citable, and authority-rich so AI models pull your insights into answers users see first.
This skill audits your current AI visibility, identifies gaps in structure and authority, and recommends optimizations that increase the likelihood your content appears in AI-generated responses—without requiring Google rank.
When to Use This Skill
Launching a new content pillar — Before publishing a guide, benchmark AI visibility and build authority signals from day one
Competitive content analysis — Understand why competitor content ranks in AI Overviews and ChatGPT summaries, then out-structure them
Refreshing underperforming content — Content with good Google traffic but zero AI citations needs extractability and authority tuning
Building thought leadership — Ensure your expert research and original insights surface in AI-cited sources, not as generic summaries
Planning content strategy — Use Harmony conversation data (sales calls, customer interviews) to identify topics and angles AI models need more coverage on
Audit before major campaigns — Validate that web properties and third-party presences are AI-bot accessible and properly formatted
Monitoring AI citation rates — Track which content pieces appear in AI responses over time and optimize the laggards
Instructions
Step 1: Map Your Current AI Visibility
Identify 8–12 core topics or content pieces your audience cares about. Search each topic in ChatGPT, Perplexity, Google Search (look for AI Overviews), and Claude with identical prompts. Document:
Whether your content appears in cited sources
Which competitor or third-party content the AI returns instead
The exact passage the AI extracts (if yours is cited)
Position in the response (top cite vs. lower mention)
Store results in a spreadsheet with columns: Topic | Your Content | Competitor 1 | Competitor 2 | AI Citation Status | Current Gap.
Step 2: Audit Content Extractability
Review your 3–5 highest-priority content pieces. For each, check:
Definition blocks: Does the piece open with a clear, one-sentence definition of the core term?
Comparison tables: For how-to or versus content, are comparisons in structured table format (not buried in prose)?
FAQ sections: Does the piece include an FAQ block with 5–8 Q&A pairs matching user intent?
Data/statistics: Are original findings or third-party stats cited with sources, author attribution, and publication date?
Subheading clarity: Can a reader skim subheadings and understand the piece's argument without reading body text?
Rate each piece on a scale (Poor | Fair | Good | Excellent) and flag 2–3 quick wins (low-effort, high-impact edits).
Medium effort (2–4 hrs): Rewrite subheadings for clarity, add author bio with credentials, cite new sources
Larger projects (1–2 days): Incorporate new research or case study data, restructure guide around AI-friendly sections
Prioritize optimizations that address authority gaps first (customers cite you less because credibility is unclear), then extractability (content can't be parsed by AI).
Step 7: Set Up Citation Monitoring
Establish a monthly or quarterly check:
Re-run searches from Step 1 using identical prompts
Compare citation frequency month-over-month
Flag which optimizations moved the needle
Identify new competitor content entering AI responses
Document baseline metrics before and after optimization.
Output Format
Create a comprehensive audit report (30–50 lines minimum) with these sections:
AI SEARCH OPTIMIZATION AUDIT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
EXECUTIVE SUMMARY
Current AI Visibility Score: [X/10]
Key Finding: [One sentence summary of biggest gap or opportunity]
Recommended Priority: [Quick Wins / Authority / Extractability]
CURRENT AI VISIBILITY MAP
Topic | Content Title | Cited in ChatGPT? | Cited in Perplexity? | Cited in GEO? | Primary Competitor | Gap Assessment
[Fill with 8–12 rows of data]
CONTENT EXTRACTABILITY AUDIT
Content Piece: [Title]
Definition Block: [Poor | Fair | Good | Excellent] — Notes
Comparison Tables: [Rating] — Notes
FAQ Coverage: [Rating] — Notes
Author Attribution: [Rating] — Notes
Data & Sources: [Rating] — Notes
Subheading Clarity: [Rating] — Notes
Overall Extractability: [Rating]
Quick Wins: [2–3 specific edits, estimated effort]
[Repeat for 3–5 pieces]
AUTHORITY SIGNAL ASSESSMENT
✓ Existing Strengths:
- [Signal 1] — Why this matters for AI
- [Signal 2]
- [Signal 3]
✗ Gaps (Blocking AI Citations):
- [Gap 1] — Impact on citations
- [Gap 2]
- [Gap 3]
AI BOT ACCESS VERIFICATION
Robots.txt Status: [Allows AI crawlers: Yes | No | Partial]
Noindex Tags: [Present on: (list pages)]
Paywall/Login Barriers: [Present on: (list pages)]
Accessibility Score: [Green | Yellow | Red]
STRUCTURE OPTIMIZATION ROADMAP
Quick Wins (Week 1):
1. [Specific edit, estimated time]
2. [Specific edit, estimated time]
3. [Specific edit, estimated time]
Medium Effort (Weeks 2–3):
1. [Structural change, effort estimate]
2. [Structural change, effort estimate]
Larger Projects (Month 1):
1. [New research, data integration, or major rewrite]
CITATION MONITORING BASELINE
[Include current citation counts before optimization]
NEXT STEPS
- [ ] Schedule content update planning meeting
- [ ] Brief product/design on schema markup needs (if applicable)
- [ ] Add to content calendar: [X updates across Y pieces]
- [ ] Set reminder for 60-day re-audit
Example
AUDIT: Meridian Analytics — "What is predictive analytics?" Content Pillar
EXECUTIVE SUMMARY
Current AI Visibility Score: 4/10
Key Finding: Competitors dominate AI responses for predictive analytics queries, but Meridian's guides lack clear definitions and third-party mentions that AI models cite.
Recommended Priority: Authority signals + Extractability
CURRENT AI VISIBILITY MAP
Topic
Content
ChatGPT
Perplexity
GEO
Competitor
Gap
What is predictive analytics?
Meridian guide (2024)
No
Mentioned 2x
No
Gartner, IBM
Low authority signal; definition buried in intro
Predictive analytics vs. descriptive
Meridian comparison (2023)
No
No
No
Microsoft, DataCamp
Competitor content better structured as table
Predictive analytics for retail
Meridian case study (2024)
No
No
No
Harvard Business Review, McKinsey
Meridian data too internal; needs third-party quotes
Time series forecasting
No Meridian content
—
—
—
Statsmodels, Kaggle
Gap: No Meridian perspective
Demand forecasting tools
Meridian product guide
No
Partial (1 cite)
No
Gartner Magic Quadrant
Product guide lacks independent validation
CONTENT EXTRACTABILITY AUDIT
Meridian Guide: "What is Predictive Analytics?" (2024)
Definition Block: Fair — Opens with history, not definition. AI models struggle to extract core concept in first two sentences.
Quick Win: Rewrite first paragraph to start with: "Predictive analytics is the practice of using historical data, statistical models, and machine learning to forecast future outcomes."
Comparison Tables: Poor — "Differences between predictive and descriptive" buried in prose across 400 words.
Medium Effort: Extract into 4-column table (Dimension | Predictive | Descriptive | When to Use).
FAQ Coverage: Good — 6 Q&As included, but positioned at bottom (AI models may truncate before reaching).
Author Attribution: Fair — Author name listed, no title or credentials.
Quick Win: Add one-line bio: "Sarah Chen, Data Science Lead at Meridian, 8+ years in predictive modeling."
Data & Sources: Poor — Original research mentioned (500+ customer implementations) but not formally cited.
Medium Effort: Create callout box: "Meridian analyzed 500+ customer implementations (2023–2024) and found predictive models improved forecast accuracy by 34% on average."
Subheading Clarity: Good — Subheadings are topic-focused, but some are passive.
Quick Win: Change "How predictive analytics is used" to "Five ways predictive analytics improves business decisions."
Content Behind Paywalls or Login Walls: AI models can't extract what they can't access. If your best research is gated, consider an open-access summary or teaser on your public domain with full version behind login. Monitor whether AI citations increase after removing friction.
Niche or Technical Topics: Smaller audiences mean fewer web links and thus fewer training examples for AI models. Compensate by building authority on adjacent, higher-traffic topics, then linking to niche content. Cite yourself.
New Company or Domain: If your domain is under 6 months old, AI models may weight you lower (low backlink history). Prioritize third-party placements (guest posts, Reddit, Twitter/X mentions) to build off-domain authority faster than SEO.
Real-Time or News-Driven Content: AI models trained on snapshots of the web may lag current events. If your content is timely, include publication date, last-updated timestamp, and cite current sources clearly. Monitor AI responses to trending queries.
Multi-Language or Regional Content: Some AI models handle non-English content differently. If you serve multiple regions, audit each language version separately. Check robots.txt doesn't inadvertently block regional variants.
Competitive Pricing or Proprietary Models: If you can't share exact pricing or methodology due to competition, focus on authority and third-party validation instead (customer wins, analyst reports, media mentions). AI models cite these signals even if product details are closed.
Harmony Integration Notes
This skill pulls power from Harmony conversation data:
Sales calls: Customer language in pitches reveals how AI models should phrase your insights.
Customer interviews: Discovery calls highlight gaps or angles competitors miss; feed these into content strategy.
Slack discussions: Team conversations about customer wins, complaints, or competitive positioning surface topics AI models need.
Email threads: Customer questions in email reveal recurring pain points and authority angles content should address.
Link audit findings back to conversation transcripts. If Harmony data shows customers often ask "How does X differ from Y?" but your content lacks a comparison, that's a high-confidence optimization signal.
