ai-seo

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• AI SEO
• answer engine optimization
• optimize for ChatGPT
• optimize for Perplexity
• how do I show up in AI answers

What it does

ai-seo is a Claude Code skill from Corey Haines’s marketing-skills repo. It turns Claude into an answer-engine optimization specialist whose job is making your content extractable and citable by AI systems — not just rankable in traditional search. The skill activates when you mention “AI SEO”, “AEO”, “GEO”, or “optimize for AI Overviews”, and walks through visibility audit, content structure, authority signals, and machine-readable files.

The output of a session is a prioritized AI-visibility plan: a query-by-platform citation matrix showing where you appear and where competitors do, an extractability checklist per priority page, robots.txt review for the AI crawlers (GPTBot, PerplexityBot, ClaudeBot, Google-Extended), and recommendations for /llms.txt and /pricing.md machine-readable surfaces.

When to use it

Reach for it when:

• You’ve noticed traffic from AI sources but don’t know which pages drive citations
• Competitors are appearing in AI Overviews for queries you used to rank for
• You’re publishing content and want it structured for AI extraction from day one

When not to reach for it:

• Your traditional SEO foundation isn’t there yet — this layer assumes it
• You’ve blocked AI crawlers as a policy decision; the skill won’t override that

Install

The skill is distributed via Corey Haines’s marketing-skills repo. Install via the repo’s recommended path — copy the ai-seo SKILL.md into your project’s .claude/skills/ai-seo/ directory, or use the repo’s plugin install if you’ve set it up.

Once installed, the skill activates on the trigger phrases above. The first time it runs, it will check for .agents/product-marketing-context.md (or .claude/product-marketing-context.md) — populating that file with your product context first dramatically improves output quality across all of Haines’s marketing skills.

What a session looks like

A typical session has three phases:

1. Visibility audit. Pick 10–20 priority queries. For each, check whether AI Overviews, ChatGPT, and Perplexity surface an answer, and who’s cited. Gaps become the work queue.
2. Three-pillar optimization. Structure (extractable answer blocks, FAQ schema, comparison tables), Authority (citations, statistics, expert attribution), Presence (Wikipedia, Reddit, review sites — third-party citations matter as much as your own domain).
3. Machine-readable surfaces. The skill closes by recommending /llms.txt, /pricing.md, and a robots.txt that allows GPTBot, PerplexityBot, ClaudeBot, and Google-Extended.

The discipline that makes it work: AI systems extract passages, not pages. Every key claim has to work as a standalone 40–60-word block, ideally with a numeric proof point and a source.

Receipts

Honest reporting on what ai-seo produces and where it has limits:

Where it works well:

• Catches the “all my competitors get cited and I don’t” pattern by inspecting actual AI answers, not guessing
• Turns the abstract “make it AI-friendly” instruction into specific block-level rewrites
• The robots.txt review surfaces accidentally-blocked AI bots that quietly suppress citations

Where it backfires:

• AI citation patterns shift fast; what works this quarter may not next quarter
• Rewriting a page for AI without checking it still reads well for humans is a real risk — the skill warns about this but won’t catch every regression

Pattern that works: trigger on a page you already rank for in traditional search. AI citation correlates strongly with rank, so the structural rewrite compounds rather than starting from zero.

Source and attribution

Originally written by Corey Haines. The canonical SKILL.md and any supporting files live in the ai-seo folder of his marketing-skills repository.

License: MIT. You can install, adapt, and redistribute the skill, with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and any updates, defer to the source repo.