SEO content strategy for the AI Overviews era (2026). Research keywords, analyze SERP + AI citations, generate blog posts optimized for both Google ranking AND AI citation. Handles keyword research, competitor gap analysis, content briefs, full article generation with schema markup, and AI-citation-optimized structure. Use when asked to write blog posts, do keyword research, create content briefs, optimize for SEO, improve search rankings, get cited by AI, or build topic cluster authority.
SEO has changed. AI Overviews appear on 48% of Google queries. Getting cited by AI systems (Google AIO, ChatGPT, Perplexity) is now as important as ranking in the top 10. This skill handles both.
User says "write a blog post about X" or "do keyword research for X":
Use web_search to find keyword opportunities:
web_search: "[topic] site:ahrefs.com OR site:semrush.com keyword difficulty"
web_search: "people also ask [topic]"
web_search: "[topic]" (examine autocomplete suggestions)
Evaluate each keyword on:
Output a keyword map: primary keyword + 5-10 secondary/LSI keywords + intent classification.
For detailed keyword research methodology, see references/keyword-research.md.
For the primary keyword, analyze what currently ranks:
Key metrics to extract:
For the AI citation analysis framework, see references/ai-citation.md.
Generate a brief containing:
# Content Brief: [Title]
**Primary keyword:** [keyword]
**Secondary keywords:** [list]
**Search intent:** [informational/transactional/etc.]
**Target word count:** [based on competitor analysis, typically 2000-4000]
**AI Overview status:** [present/absent for this query]
## Required Sections
- [H2 headings based on competitor analysis + gap fill]
## Questions to Answer
- [From "People Also Ask" + competitor gaps]
## Differentiation Angle
- [What we cover that competitors don't]
## Internal Links
- [Other pages on the site to link to/from]
## Citation Optimization Notes
- [Specific stats, data, or claims to include for AI citation]
Follow this structure for maximum ranking + AI citation potential:
Lead with a direct answer (40-60 words) before any elaboration. AI systems extract the first substantive paragraph.
Use clear H2/H3 hierarchy matching search intent. Each H2 section should be self-contained (134-167 words) -- this matches the AI extraction window.
Include stats with source attribution every 150-200 words. AI systems cite data-rich content. Format: "According to [Source], [specific stat]."
Add a summary table or comparison if applicable. AI systems frequently cite tabular data.
Answer "People Also Ask" questions as H2 sections with concise 2-3 sentence answers followed by elaboration.
Keep paragraphs to 2-3 sentences max. Shorter paragraphs = easier AI extraction.
Include an FAQ section with JSON-LD FAQ schema.
Title: [Primary keyword] -- [benefit or year] (50-60 chars)
Meta description: [Direct answer to query + CTA] (150-160 chars)
For the complete writing checklist, see references/writing-checklist.md.
Generate JSON-LD for:
See references/schema-templates.md for copy-paste JSON-LD templates.
Single pages rank worse than topic clusters in 2026. Build clusters:
Pillar page: "Complete Guide to [Topic]" (3000-5000 words)
|-- Cluster: "[Topic] for beginners" (2000 words)
|-- Cluster: "[Topic] vs [Alternative]" (2000 words)
|-- Cluster: "Best [Topic] tools/tips" (2000 words)
|-- Cluster: "[Topic] common mistakes" (1500 words)
|-- Cluster: "[Topic] advanced guide" (2500 words)
Each cluster page links to the pillar and to other cluster pages. This builds topical authority that AI systems use for entity confidence scoring.
Beyond traditional SEO, optimize for AI search engines: