GEO Tactics Complete Guide — The Full Implementation Map

Executive Summary

• GEO is composed of five main tactics: llms.txt / Schema.org structured data / FAQ pages / AI crawler accommodation / internal link strategy.
• In GEO Meter's smoke analysis (2 topics x roughly 20 domains), llms.txt deployment in particular showed a +20-30pp gap in implementation rate between the top and bottom AI citation groups, suggesting that implementing fundamental tactics directly drives citation acquisition.
• The recommended order of attack is "observe → llms.txt → Schema.org → FAQ → crawler → internal links" — arranged from lower cost / higher impact downward.
• You don't have to do everything at once. The realistic approach is to proceed one tactic at a time, measuring impact as you go.

1. The Big Picture of GEO Tactics

The five main tactics to work on for GEO are:

→ In cost-performance order: "llms.txt → crawler → Schema → FAQ → internal links."

Impact x Implementation Difficulty Matrix

Placing each tactic by "impact" and "implementation difficulty" makes priorities visible at a glance:

Recommended Order of Attack

We'll explain each tactic in order.

2. Deploy llms.txt (Highest Priority)

llms.txt is a file that uses Markdown to explicitly indicate your site's main content for AI crawlers.

How to Deploy

Place /llms.txt at the site root. Minimal structure:

# Site name

> One- to two-line description of the site

## Main pages

- [Page name](URL): brief description

Impact (GEO Meter Smoke Analysis: Otaku Travel Topic, Roughly 20 Domains)

A different topic (fat reduction) also showed a +20pp gap, confirming a consistent tendency across multiple topics.

Implementation Tips

• Markdown format (not HTML)
• Not too long (rough guide: 2,000-5,000 characters, with emphasis on summary)
• Regular updates (around once a month; explicitly noting a lastmod-style date is recommended)

3. Schema.org / JSON-LD Structured Data

Implement Schema.org in JSON-LD format. Think of it as "a vocabulary that tells AI what's on the web page."

The Four Most Important Schemas

Implementation Example (Article Schema)

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "What is GEO?",
  "datePublished": "2026-05-24",
  "dateModified": "2026-05-24",
  "author": {
    "@type": "Organization",
    "name": "XX Editorial Team"
  },
  "publisher": {
    "@type": "Organization",
    "name": "XX Inc."
  }
}
</script>

Impact (Position in the GEO Context)

Implementing Schema.org is the most important foundational measure for AI to understand "what this page represents" in a machine-readable form. It tends to raise priority as an AI citation candidate and also has a secondary impact on SEO rich results.

Common Mistakes

• Writing it as HTML microdata (JSON-LD is recommended)
• Missing required fields (datePublished, author, etc.)
• Multiple instances of the same schema (Article should appear once per page)

4. Designing FAQ Pages

Q&A-format pages with FAQPage Schema are the structure most likely to be cited in AI search.

Why FAQs Work

AI search engines (Perplexity / SearchGPT / Gemini / Claude) generate "direct answers" to user questions. The FAQ format is already structured as "question → answer," making it the easiest format for AI to handle.

Impact (Position in the GEO Context)

FAQ pages perfectly match the structure of "direct answers to questions" that AI search wants, making them one of the structures most likely to be cited. Combining with FAQPage Schema maximizes impact.

Good FAQs vs. Bad FAQs

Rough Guide for Getting Started

• Begin with 5-10 representative Q&As
• Add 2-3 new Q&As per month
• Reflect questions that come up frequently in customer support

5. AI Crawler Accommodation (robots.txt)

Explicitly Allow AI crawlers (GPTBot / Claude-Web / Google-Extended / PerplexityBot) in your robots.txt.

Example Configuration (GEO Recommended)

# GPTBot (ChatGPT / SearchGPT) User-agent: GPTBot Allow: / # Claude-Web (Claude) User-agent: Claude-Web Allow: / # Google-Extended (Gemini / AI Overview) User-agent: Google-Extended Allow: / # PerplexityBot (Perplexity) User-agent: PerplexityBot Allow: / # Make the sitemap explicit too Sitemap: https://example.com/sitemap.xml

Why You Should Allow

• Disallowing completely removes you from AI citation candidates
• Older robots.txt templates may block AI crawlers (needs review)
• The feeling of "I don't want my content used as training data" is understandable, but you simultaneously lose citation opportunities

How to Verify

curl https://your-site.com/robots.txt

Check that Disallow: / is not set against AI crawlers.

6. Internal Link Strategy

Systematize internal links to build topical authority across your entire site.

Topic Cluster Structure

• Pillar Page: a complete guide to a specific theme (3,000-5,000 characters)
• Cluster Page: an individual deep-dive article (1,500-3,000 characters), linking to the Pillar
• Glossary: term collection, linked from every article

Anchor Text Optimization

Instead of vague links like "here" or "click here for details," use keywords that represent the linked content as the anchor text.

<!-- Bad example -->
For details, <a href="/llms-txt-guide">click here</a>

<!-- Good example -->
See <a href="/llms-txt-guide">how to write llms.txt correctly</a>

7. Implementation Priority

The recommended order of attack:

Rough Time Estimates per Step

8. By Case: Your Personal Plan of Attack

Case 1: B2B SaaS with Engineers On Staff and Nothing Done Yet

Week 1: Deploy llms.txt (1-2 hours) + verify AI crawler Allow in robots.txt

# robots.txt example User-agent: GPTBot User-agent: Claude-Web User-agent: Google-Extended User-agent: PerplexityBot Allow: / Sitemap: https://example.com/sitemap.xml

Weeks 2-3: Implement Schema.org (Organization + Article)

Week 4 onward: Start with 5 FAQ entries, add 2-3 per month

Case 2: SMB Outsourcing to a Web Production Agency

1. Grasp the current state with the free diagnostic → share with the production agency
2. Ask them to "deploy llms.txt" (market price 20,000-30,000 yen)
3. Ask for Schema.org JSON-LD to be added to your existing site's <head>
4. Add FAQPage Schema to your existing FAQ pages

→ Initial cost including all outsourcing comes out to around 100,000-300,000 yen.

Case 3: A Company Already Investing in SEO

1. Verify existing Schema.org implementation (Article / Product, etc.)
2. Check whether AI crawlers (such as GPTBot) are Allowed in robots.txt (they're surprisingly often Disallowed)
3. Add llms.txt
4. Add FAQPage Schema to existing FAQs

→ Layer GEO-specific tactics on top while maximizing the use of SEO assets.

Case 4: News / Media Site

1. Apply Article Schema to every article (author / publish date / update date)
2. Author Schema (author profiles)
3. List main series and tag pages in llms.txt
4. Publish proprietary data with awareness of being a cited source

9. Frequently Asked Questions

Q. If I commission an SEO agency, will GEO be covered too?

Partially yes (Schema.org / internal links and the like work for both). However, llms.txt deployment and AI crawler accommodation are often not in the standard menu of SEO agencies. You need to explicitly request "GEO measures."

Q. Can WordPress sites do this too?

Yes. Schema.org JSON-LD can be implemented via plugins (such as Yoast SEO or RankMath). For llms.txt, you can edit the theme or place it directly on the server. FAQ pages are the same.

Q. What about sites that render dynamically with JavaScript?

Some AI crawlers don't execute JavaScript (GPTBot has limited support). HTML delivery via SSR / SSG is essential for GEO measures. SSR configurations with Next.js / Nuxt and similar are recommended.

Q. How long until you see results?

It varies by tactic:

• llms.txt + robots.txt: 1-2 weeks until AI crawler access changes
• Schema.org implementation: 1-2 months until AI citation behavior changes
• FAQs + internal links: 3-6 months of continuous effort to establish topical authority

Q. How do I measure impact?

• Measure it yourself: monthly manual queries against ChatGPT / Claude / Gemini → record citation sources
• Use a tool: automated observation via GEO Meter (monthly SOV / citation count)
• Indirect indicators: track inbound traffic from AI-class user agents in Google Search Console

10. Summary

• GEO consists of five main tactics: llms.txt / Schema.org / FAQ / crawler / internal links
• The order of attack is "observe → llms.txt → Schema → FAQ → crawler → internal links"
• In GEO Meter's smoke analysis (2 topics x roughly 20 domains), there is a strong correlation between the implementation rates of each tactic and AI citation top companies
• Priority differs by industry and situation: phase it according to your case
• Confirm impact via monthly reviews, and revisit priorities as needed

Related Resources

• GEO Fundamentals Complete Guide — Pillar 1
• GEO Management Complete Guide — Pillar 3
• Glossary — detailed explanations of specialized terms for each tactic
• Latest Articles — monthly reports and industry trends
• GEO Meter Free Diagnostic — measure your current state in 3 minutes