| name | ai-seo-optimization |
| description | Audit and improve traditional search ranking: on-page SEO, schema markup, entity recognition, and content structure. Use for "audit my SEO", "improve search rankings", "implement schema markup", or "fix my site structure". For AI-visibility auditing and agentfacts, use geo-optimizer. |
| version | 0.1.1 |
AI SEO Optimization Skill
Modern SEO for the AI-powered search era. Covers entity-based optimization, AI citation building, multi-platform visibility, E-E-A-T implementation, and technical SEO for 2025 and beyond.
AI SEO Audit Process
Run this sequence for any SEO analysis task:
- Entity Analysis — Check brand recognition across AI platforms (Google Knowledge Graph, Wikidata, Wikipedia)
- Content Structure — Assess AI-readability: clear headings, tables, FAQs, direct answers at top
- Schema Coverage — Audit structured data completeness (Organization, Article, FAQ, HowTo, Product)
- Citation Tracking — Check visibility in AI responses (Google AI Overviews, ChatGPT, Perplexity)
- Technical Health — Core Web Vitals, mobile-first, robots.txt, XML sitemaps
Entity-Based SEO
Brand Entity Building:
- Establish unique, consistent brand name and description across all platforms
- Create and verify Wikipedia/Wikidata entries where notability exists
- Build social profile network (Twitter/X, LinkedIn, GitHub, Crunchbase) with consistent brand data
- Implement
sameAs schema markup linking all brand entity sources
Schema Priority Order:
{
"@context": "https://schema.org",
"@graph": [
{ "@type": "Organization", "name": "...", "sameAs": ["..."] },
{ "@type": "WebSite", "publisher": {"@id": "#organization"} },
{ "@type": "Article", "author": {"@type": "Person"}, "dateModified": "2025-01-01" }
]
}
AI Citation Optimization
Content structure for AI platforms to cite:
- Quick Answer at top — Direct 2-3 sentence response to the main question before any detail
- Structured sections — Clear H2/H3 hierarchy with descriptive headings
- Data tables — Comparisons and specifications in table format (AI systems love to pull these)
- FAQ sections — Q&A format at the bottom of long-form content
- Authoritative sources — External citations, author credentials, publication dates
Platform-Specific Requirements:
| Platform | Key Factor | Optimization |
|---|
| Google AI Overviews | Featured snippet presence | Direct answer + FAQ schema |
| ChatGPT | Training data coverage + recency | Authoritative content + consistent publishing |
| Perplexity | Source credibility + freshness | High-quality external links + recent dates |
| Gemini | Technical accuracy + depth | Expert author bios + citations |
| Bing Chat | Microsoft ecosystem | Bing Webmaster Tools + IndexNow |
Content Optimization Workflow
- Query Analysis — Identify natural language and long-tail variations of the target topic
- Topic Clustering — Map all subtopics and create comprehensive coverage across linked pages
- Internal Linking — Build semantic relationships between related content
- Format Variety — Use tables, numbered lists, Q&As, and comparison charts (all AI-parseable formats)
- Human Voice — Add first-person experience, original data, and brand-specific perspective
Striking-Distance Optimization (Google Search Console)
Striking distance is the band of queries where a page already ranks in position 4-20 in Google — the bottom of page 1 through the top of page 2. Google already treats these pages as relevant enough to rank, and the click-through-rate curve is steep near the top: clicks run roughly 25-30% at position 1 and drop to low single digits by position 10, so a small move up captures most of the traffic still on the table. A query moving from position 8 to position 3 can multiply its clicks several times over, often from a title tag rewrite or a handful of added internal links. Because the ranking signal already exists, these queries are usually the cheapest wins available on a site that already gets some traffic.
Workflow:
-
Pull GSC performance data for the property, at the query + page level, over a recent window (28-90 days is typical). Whether exporting from the Search Console UI's Performance report or calling the Search Console API's searchanalytics.query method, the required fields are the same: query, page, clicks, impressions, ctr, position.
-
Filter to the striking-distance band — average position between 4 and 20 — and drop rows with negligible impressions. The exact impression floor depends on the property's traffic volume: a low-traffic site may need a floor as low as 10, a high-traffic one several hundred.
-
Score each query by Impact × Confidence:
- Impact — impressions, or an estimate of clicks recoverable if the query moved into position 1-3, using the CTR curve above
- Confidence — a proximity-to-top multiplier: a query at position 5 sits closer to the CTR cliff than one at position 18, so it scores higher
Multiply the two and rank descending. The result surfaces queries with meaningful search volume and only a short climb left to the top three positions.
-
Output a ranked action list of page/query pairs, each with its current position, impressions, click estimate, and priority score, ready to hand to a content or on-page optimization pass.
Data source note: this technique depends on the property's own Google Search Console data — real rankings and click behavior pulled directly from Google for that specific site. Generic keyword-research tools estimate volume and difficulty from external indexes; only Search Console holds a site's actual position and click history for that property. Access the data via the GSC UI export or the Search Console API, and verify the current endpoint name and field set against Google's own documentation before implementing — the field list above (query, page, clicks, impressions, ctr, position) is the contract to build against.
Striking-distance methodology adapted from the MIT-licensed ericosiu/ai-marketing-skills seo-ops skill.
E-E-A-T Implementation
Experience: Demonstrate first-hand knowledge — case studies, original research, specific examples from actual use
Expertise: Author credentials in bio, bylines on all content, topic specialization signals
Authoritativeness: Industry citations, press mentions, guest contributions to authority publications
Trustworthiness: Transparent sourcing, fact-checking processes, clear contact information, privacy policy
Notability: Brand mentions across independent sources, recognition from authorities in the field
Technical SEO
Core Web Vitals targets:
- LCP (Largest Contentful Paint): < 2.5s
- INP (Interaction to Next Paint): < 200ms
- CLS (Cumulative Layout Shift): < 0.1
Crawl and index health:
- XML sitemap with priority and lastmod signals
- robots.txt with explicit AI crawler rules (note: blocking AI crawlers affects training inclusion)
- Canonical tags to prevent duplicate content
- Hreflang for international sites
AI Crawler Management (robots.txt):
# Allow all crawlers including AI
User-agent: *
Allow: /
# Or selectively block AI training crawlers
User-agent: GPTBot
Disallow: /
User-agent: CCBot
Disallow: /
Key AI SEO Metrics
| Metric | What It Measures |
|---|
| Citation Share | % of AI responses mentioning brand for target queries |
| Entity Score | Knowledge graph completeness (Wikidata + schema coverage) |
| AI Visibility Index | Cross-platform presence in AI-generated responses |
| Query Coverage | % of natural language query variants ranking |
| Featured Snippet Share | % of target queries where content appears in featured snippet |
Reference Files
For detailed implementation guidance, consult:
- Invoke
Skill(marketing-skills:seo-audit) for full SEO audit workflow
- Invoke
Skill(marketing-skills:schema) for structured data implementation
- Invoke
Skill(marketing-skills:programmatic-seo) for programmatic SEO strategies