Ai Seo

AI Visibility & AI SEO Audit — Phase 14

Executive Summary

AI search is the fastest-growing traffic channel in 2025. Google AI Overviews appear for 20–35% of local service queries (SparkToro 2025). ChatGPT Search has 200M+ monthly active users. Perplexity processes 100M+ queries/month. Being cited in AI answers = significant visibility gain; being absent = invisible to AI-first searchers. This phase audits where the client appears (or doesn't) across all major AI platforms, identifies content gaps causing AI invisibility, and produces a prioritized action plan for AI citation dominance.

2025 AI search benchmarks:

• AIO citation rate: pages with FAQPage schema cited 3.2× more than pages without (Amsive 2025)
• ChatGPT freshness: pages updated within 30 days cited at 76.4% rate vs. 31.2% for 90+ day old pages
• Perplexity local query share: 18% of all queries are local business intent (Perplexity data, 2025)
• AIO + organic position 1 = ~45% CTR combined (vs. ~25% for position 1 alone when AIO shows)

---

Tools for This Phase

| Tool | Purpose | Cost | |------|---------|------| | Google Search (incognito) | Test AIO presence for top 20–30 target keywords | Free | | ChatGPT (GPT-4o with Search) | LLM visibility test — local service queries | Free/Paid | | Perplexity | AI search citations for local service queries | Free | | Google Gemini | Gemini AI visibility test + source attribution | Free | | Microsoft Copilot | Bing-powered AI visibility | Free | | Google Rich Results Test | FAQPage, HowTo, Speakable schema validation | Free | | Ahrefs | Featured snippet ownership — AIO pulls from snippets | Paid | | SEMrush | Featured snippet tracking per target keyword | Paid | | AlsoAsked.com | PAA question mapping — AIO uses PAA patterns | Freemium | | Otterly.ai | Track AIO mentions over time (monitoring) | Paid |

---

The AI Search Landscape (2025-2026)

AI search is no longer emerging — it IS the mainstream:

• Google AI Overviews: Shown for ~20-30% of queries (growing)
• Google AI Mode: Full conversational search experience
• ChatGPT Search: Now default for 200M+ users with web access
• Perplexity: 100M+ monthly queries, growing fast
• Microsoft Copilot: Integrated into Windows/Edge/Bing
• Gemini (Google): Deep integration with Android + Google Workspace
• Claude (Anthropic): Growing presence with web search

For local businesses, AI search drives both:

1. Direct visibility (being mentioned in AI answers)
2. Competitive displacement (competitor gets the mention instead)

---

Step 1: Read Project Context

Read {AUDIT_DIR}/intake-data.md — business name, URL, services, location. Read {AUDIT_DIR}/competitor-profiles.md — competitor AI visibility signals. Read {AUDIT_DIR}/onpage-findings.md — content structure gaps.

---

Section 1: Google AI Overviews (AIO) Audit

Test Protocol

Test top 20-30 target keywords in Google Search (incognito, desktop + mobile):

For each query, document: | Query | AIO Shows? | Client Cited? | Competitors Cited? | Source URLs Used | AIO Format | |-------|-----------|--------------|-------------------|-----------------|-----------| | [query] | Yes/No | Yes/No | [names] | [URLs] | Paragraph/List/Table/Steps |

AIO Appearance Patterns to Identify

• Which query types trigger AIO? (Informational, local, how-to, comparison?)
• Which content format is used in AIO? (Paragraphs, numbered steps, bullet lists?)
• What's the word count of the cited answer section?
• Does AIO pull from the first paragraph? A specific H2 section?
• Are any client pages cited? If so, which ones and why?

AIO Optimization Strategy

To be cited in AI Overviews, content needs:

1. Direct answers — First 40-60 words answer the query directly
2. Question-format headers — H2: "What is [topic]?" / "How does [process] work?"
3. Structured formatting — numbered steps, bullet lists, tables rank well
4. Unique data/insights — AI cites specific facts, statistics, and expert opinions
5. E-E-A-T signals — Strong author credentials, institutional trust
6. Comprehensive coverage — Answer main query + related sub-questions
7. Page authority — Higher authority pages cited more often

---

Section 2: Google AI Mode Audit

Google AI Mode (launched 2024, expanding 2025) enables fully conversational search.

Test AI Mode Queries

• Enable AI Mode (Google account > Experiments or rolled-out region)
• Test conversational queries:
• "What's the best [service] in [city] and why?"
• "Compare [client] vs. [competitor 1] for [service]"
• "I need [service] near [neighborhood] — any recommendations?"
• Follow-up: "What makes them stand out?"

Document Findings

| Query | Client Mentioned? | Accuracy | Competitor Priority | Source | |-------|-----------------|---------|-------------------|--------|

---

Section 3: LLM Visibility Audit (ChatGPT, Perplexity, Gemini, Copilot, Claude)

Test Matrix

For each platform × query type:

Platforms to test:

• ChatGPT (GPT-4o with Search)
• Perplexity
• Google Gemini
• Microsoft Copilot
• Claude.ai (Anthropic)

Query types to test:

1. Best [service] in [city]
2. [Service] near [neighborhood/landmark]
3. [Business Name] — accuracy check
4. [Business Name] reviews — reputation
5. How to find a good [service type] in [city]
6. [Specific service] cost in [city]

For each result document:

• Business mentioned? (Yes/No)
• Accuracy of info (address, phone, hours, services)?
• Positive/negative framing?
• Competitors mentioned instead?
• Source URLs cited?
• Position (1st mention / 2nd / not mentioned)?

LLM Visibility Score (per platform)

• Mentioned accurately in top 3 answers: 5/5
• Mentioned but with errors: 3/5
• Mentioned peripherally: 1/5
• Not mentioned: 0/5

---

Section 4: AEO — Answer Engine Optimization Assessment

AEO focuses on structuring content so AI can extract and use it as answers.

Content Assessment

For each key service page and FAQ page:

• [ ] Does each section start with a direct, standalone answer?
• [ ] Are definitions provided for key service terms?
• [ ] Are process steps numbered and clearly labeled?
• [ ] Are comparisons structured in tables?
• [ ] Are FAQs in explicit question + answer format (not buried)?
• [ ] Are specific figures, statistics, and data points included?
• [ ] Can the answer be extracted and read without surrounding context?

FAQ Content Audit

• FAQ page exists?
• FAQs based on real customer questions (not generic)?
• FAQPage schema implemented?
• PAA (People Also Ask) boxes captured with existing content?
• Test: do current FAQs appear in PAA for target queries?

Featured Snippet Optimization

Since AI Overviews pull from featured snippets:

• Which target queries have featured snippets?
• Does the client own any?
• What format does the winning snippet use?
• What page structure modifications are needed to win it?

---

Section 5: GEO — Generative Engine Optimization Assessment

GEO focuses on making content easily citable by AI systems.

Citeability Checklist

• [ ] Specific, quotable statements with data?

Good: "Chicago homeowners saved an average of $2,400/year after our insulation upgrade" Bad: "We help homeowners save money on energy bills"

• [ ] Original research, surveys, or case studies?
• [ ] Named expert quotes with credentials?
• [ ] Step-by-step instructions with clear numbered format?
• [ ] Comparison tables with specific data?
• [ ] Local statistics about the service area?
• [ ] Industry benchmarks stated clearly?

Source Authority Assessment

AI systems cite authoritative sources. Check:

• Domain authority and trust indicators
• E-E-A-T signals present (author, expertise, experience demonstrated)
• Linked to from other authoritative sources in the niche?
• Consistently mentioned on industry resource pages?

Entity Optimization for AI

AI systems understand entities better than raw keyword text:

• Is the business entity clearly defined (who, what, where, when)?
• Are service entities named consistently (using industry-standard terminology)?
• Are geographic entities precise (neighborhood, city, region)?
• Do related entities connect? (business → services → location → team)?
• Schema markup reinforces all entity relationships?

---

Section 6: Structured Data for AI Readability

Schema types most valuable for AI citation: | Schema Type | AI Benefit | |------------|-----------| | LocalBusiness | Business entity establishment | | FAQPage | Direct Q&A extraction | | HowTo | Step-by-step answer extraction | | Article + Author | E-E-A-T, content authority | | AggregateRating | Trust signal for AI recommendations | | Speakable | Marks content optimized for voice/AI reading | | Service | Clear service entity definition | | Review | Specific review content citation |

Validate all schema at: search.google.com/test/rich-results

---

Section 7: Competitor AI Visibility Comparison

| Platform | Client Score | Comp 1 | Comp 2 | Comp 3 | |----------|-------------|--------|--------|--------| | Google AIO | | | | | | ChatGPT Search | | | | | | Perplexity | | | | | | Gemini | | | | | | Copilot | | | | | | Overall AI Score | | | | |

Findings: Which competitor dominates AI visibility and why? What content do they have that the client lacks?

---

Section 8: AI SEO Action Plan

Priority Matrix

| Action | Impact (1–5) | Feasibility (1–5) | Priority | Effort | |--------|-------------|-------------------|---------|--------| | Add FAQPage schema to all service pages | 5 | 5 | 25 | 30 min/page | | Rewrite service page intros (direct 50-word answer) | 5 | 4 | 20 | 30–60 min/page | | Add specific data points (costs, timelines) to pages | 5 | 4 | 20 | 30–60 min/page | | Create dedicated FAQ page (conversational queries) | 4 | 5 | 20 | 2–4 hrs | | Add question-format H2s to top 5 service pages | 4 | 4 | 16 | 30 min/page | | Add HowTo schema to process pages | 4 | 5 | 20 | 30 min/page | | Publish original research / local statistics page | 5 | 3 | 15 | 4–8 hrs | | Build E-E-A-T author profiles (credentials, bios) | 4 | 4 | 16 | 2–4 hrs | | Refresh pages > 30 days old (freshness = AIO boost) | 4 | 5 | 20 | 30–60 min/page | | Add Speakable schema to key answer sections | 3 | 4 | 12 | 30 min |

Immediate (Week 1–2) — Quick Wins

• [ ] Add FAQPage schema to all service pages — Effort: 30 min/page — Expected: 3.2× more AIO citations
• [ ] Rewrite service page intros with direct standalone answers (first 50 words) — Effort: 30–60 min/page
• [ ] Add specific data points (pricing, timelines, statistics) to every service page — Effort: 30–60 min/page

Short-Term (Month 1)

• [ ] Create dedicated FAQ page optimized for conversational queries — Effort: 2–4 hrs
• [ ] Restructure top 5 service pages with question-format H2s — Effort: 30 min/page
• [ ] Publish original research piece (local survey, case study, cost data) — Effort: 4–8 hrs
• [ ] Add HowTo schema to all process-based pages — Effort: 30 min/page

Medium-Term (Months 2–3)

• [ ] Build E-E-A-T author profiles for all content contributors — Effort: 2–4 hrs
• [ ] Earn citations from authoritative local/industry sites — Effort: ongoing
• [ ] Create content that fills AI citation gaps vs. top competitor — Effort: 2–4 hrs/piece

---

Scoring

| Category | Weight | |----------|--------| | AI platform visibility (AIO + ChatGPT + Perplexity + Gemini + Copilot) | 20% | | AI citability scoring (passage-level extractability) | 20% | | AI crawler access (Tier 1–3 crawlers + llms.txt) | 15% | | AEO content structure (answer blocks, FAQs, featured snippets) | 15% | | Platform-specific optimization readiness | 15% | | Schema for AI readability (FAQPage, Speakable, knowsAbout) | 15% |

---

Output

Write complete findings to {AUDIT_DIR}/ai-seo-findings.md with YAML frontmatter:

---
skill: ai-visibility/ai-seo
phase: 14
date: [YYYY-MM-DD]
business: [Business Name]
url: [URL]
score: [X/100]
aio_cited: [yes|no|partial]
chatgpt_cited: [yes|no]
perplexity_cited: [yes|no]
faqpage_schema: [yes|no|partial]
aio_keywords_tested: [X]
aio_keywords_cited: [X]
---

Include:

• Score X/100 with per-category breakdown
• Full AI visibility matrix (all 5 platforms × 6 query types)
• AIO appearance analysis for top 20–30 keywords
• Competitor AI visibility comparison table
• AEO content structure checklist (per key page)
• GEO citeability checklist
• Schema for AI readability (present vs. missing per page)
• Content gap list (what's missing for AI citation)
• Priority matrix (all actions, Impact × Feasibility scored)
• 30/90-day AI SEO action plan with effort estimates

Output files:

• {AUDIT_DIR}/ai-seo-findings.md — AI visibility audit with score and citation gap analysis
• {REPORTS_DIR}/phase-14-ai-seo.pdf — auto-generated PDF after phase completes

Key consumers:

• ai-visibility/voice-search — voice and AIO share optimization signals
• audit/onpage-seo — AIO structure requirements inform on-page optimization
• cross-cutting/serp-trust-auditor — Trust & AI Readiness (T2) dimension
• output/report-generation — AI visibility section in master report

---

Section 9: AI Citability Scoring Framework

AI models cite passages that meet specific structural criteria. GEO-optimized content achieves 30–115% higher visibility in AI-generated responses (Georgia Tech / Princeton / IIT Delhi 2024). Optimal AI-cited passages are 134–167 words, self-contained, fact-rich, and answer-first.

Citability Rubric (per content block, 0–100)

| Category | Weight | What It Measures | Scoring | |----------|--------|-----------------|---------| | Answer Block Quality | 30% | Does the passage open with a direct, quotable answer? Uses "X is..." or answer-first patterns? First 40–60 words stand alone? | 90+: every section opens with 1–2 sentence answer; 50–69: answers buried mid-paragraph; H2>H3 hierarchy? Question-based headings? Short paragraphs (2–4 sentences)? Tables for comparisons? Lists for processes? | 90+: clean hierarchy + question headings + tables/lists; 50–69: some structure; Content delivery networks (CDNs) are distributed server systems that cache and serve web content from locations geographically close to end users. A CDN reduces latency by 50–70% on average by serving assets from edge servers rather than a single origin server. The three largest CDN providers as of 2025 are Cloudflare (serving approximately 20% of all websites), Amazon CloudFront, and Akamai Technologies. > — 58 words. Self-contained: yes. Facts: 3 specific data points. Definition pattern: yes.

LOW citability (score ~15): > If you've ever wondered why some websites load faster than others, the answer might surprise you. There's this amazing technology that has been around for a while now. It's changed the way we think about web performance. Let me explain how it works. > — 52 words. Self-contained: no (no topic named). Facts: 0. Definition pattern: no.

Citation Research Data

| Finding | Source | |---------|--------| | Optimal AI-cited passage length: 134–167 words | Bortolato 2025 analysis of AI Overview passages | | Definition patterns increase citation rate by 2.1× | Georgia Tech 2024 | | Adding statistics increases citation by 40% | Princeton GEO study 2024 | | Adding authority quotations increases citation by 115% in some categories | IIT Delhi 2024 | | Fluency optimization increases visibility by 30% average | Georgia Tech 2024 | | Content with source citations cited 20–25% more often by Perplexity/ChatGPT | Industry data 2025 |

AI System Citation Preferences

| AI System | Citation Preference | |-----------|-------------------| | Google AI Overviews | Concise answer blocks (40–60 words). Content already ranking in top 10. Structured formatting (tables, lists). | | ChatGPT (Search) | Explicit definitions, named sources, recent dates. Cites 2–4 sources per response. Wikipedia source

…

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: mshahiddigital
• Source: mshahiddigital/agentic-local-seo-audit
• License: MIT

Install and usage instructions live in the source repository linked above.