seo-2026（SEO 2026 搜索）

SEO 2026 -- AI-Era Content Engine

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.

Quick Start

User says "write a blog post about X" or "do keyword research for X":

1. Run keyword research (Step 1)
2. Analyze top SERP results + AI Overview citations (Step 2)
3. Generate content brief (Step 3)
4. Write the article with AI-citation-optimized structure (Step 4)
5. Generate schema markup (Step 5)

Step 1: Keyword Research

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:

• Search intent: informational, transactional, commercial, navigational
• AI Overview presence: does this query trigger an AI Overview? (search it)
• Competition: who ranks? How deep is their content?
• Topic cluster fit: does it connect to other keywords we target?

Output a keyword map: primary keyword + 5-10 secondary/LSI keywords + intent classification.

For detailed keyword research methodology, see references/keyword-research.md.

Step 2: SERP + AI Citation Analysis

For the primary keyword, analyze what currently ranks:

1. web_search the keyword -- note top 5 results
2. web_fetch top 3 results -- analyze structure, depth, word count, headings
3. Check for AI Overview -- search the query, note what gets cited
4. Identify gaps -- what do competitors miss? What questions go unanswered?

Key metrics to extract:

• Average word count of top results
• Common H2/H3 headings across competitors
• Topics covered vs. topics missing
• Citation patterns in AI Overview (if present)

For the AI citation analysis framework, see references/ai-citation.md.

Step 3: Content Brief

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]

Step 4: Write the Article

Follow this structure for maximum ranking + AI citation potential:

AI-Citation-Optimized Structure

1. Lead with a direct answer (40-60 words) before any elaboration. AI systems extract the first substantive paragraph.

1. Use clear H2/H3 hierarchy matching search intent. Each H2 section should be self-contained (134-167 words) -- this matches the AI extraction window.

1. Include stats with source attribution every 150-200 words. AI systems cite data-rich content. Format: "According to [Source], [specific stat]."

1. Add a summary table or comparison if applicable. AI systems frequently cite tabular data.

1. Answer "People Also Ask" questions as H2 sections with concise 2-3 sentence answers followed by elaboration.

1. Keep paragraphs to 2-3 sentences max. Shorter paragraphs = easier AI extraction.

1. Include an FAQ section with JSON-LD FAQ schema.

Content Quality Rules

• E-E-A-T signals: include author credentials, first-hand experience, specific examples
• Definite language: AI systems prefer "X costs $29/month" over "X may cost around $29/month"
• Entity density: mention specific tools, people, companies by name (AI citation relies on entity matching)
• Fresh data: ~50% of AI-cited content is less than 13 weeks old. Include current-year data.
• 2000+ words minimum: pages over 20K characters average 4.3x more AI citations

Meta Tags

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.

Step 5: Schema Markup

Generate JSON-LD for:

• Article schema (always)
• FAQ schema (if FAQ section exists)
• HowTo schema (if tutorial content)
• BreadcrumbList (always)

See references/schema-templates.md for copy-paste JSON-LD templates.

Topic Cluster Strategy

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.

GEO: Generative Engine Optimization

Beyond traditional SEO, optimize for AI search engines:

1. Multi-source corroboration: get your brand mentioned on 3+ independent domains (guest posts, interviews, directories)
2. Structured claims: use definite language with numbers ("reduces cost by 40%") not vague claims
3. Question-answer format: AI systems love extracting Q&A pairs
4. Update frequency: refresh content every 8-12 weeks to stay within the 13-week citation window

Common Mistakes

• Writing for keywords without checking search intent -- wrong format = no ranking
• Ignoring AI Overviews -- if AIO shows for your keyword, optimize for citation, not just ranking
• Thin content (<1000 words) -- deep content gets 4.3x more AI citations
• No internal linking -- destroys topic cluster authority
• Stale content -- AI systems prefer content < 13 weeks old
• Generic headings -- H2s should match actual search queries