ai-seo

Overview of ai-seo skill

The ai-seo skill is for teams that want content cited in AI-generated answers, not just ranked in traditional search. It is best for marketers, founders, content leads, and SEO practitioners working on visibility in Google AI Overviews, ChatGPT, Perplexity, Claude, Gemini, and similar answer engines.

What ai-seo actually helps you do

The real job-to-be-done is to diagnose why your brand or pages are missing from AI answers, then turn that into a practical content and visibility plan. The skill focuses on whether your content is:

• indexed where the platform looks
• crawlable by relevant bots
• extractable as answer-ready passages
• credible enough to be cited

This makes ai-seo more useful than a generic “optimize my article” prompt when your goal is citation, mention share, or inclusion in recommendation-style answers.

Who should install this ai-seo skill

Install this ai-seo skill if you need to:

• improve AI visibility for product, category, or comparison queries
• evaluate whether AI bots should be blocked or allowed
• adapt SEO content for answer engines and zero-click search
• create pages that are easier for LLMs to quote, summarize, and trust
• build a repeatable workflow for AI search optimization across platforms

If you mainly need a technical SEO audit or schema implementation, this is not the first skill to reach for.

What makes this skill different

The strongest differentiator is that ai-seo is not framed as “write keyword content.” It uses a clearer operating model around three practical pillars:

• Structure: make content easy to extract
• Authority: make claims easy to trust and cite
• Presence: be visible where AI systems source answers

The repo also includes two genuinely useful support references:

• references/content-patterns.md for reusable answer-engine content blocks
• references/platform-ranking-factors.md for platform-by-platform sourcing behavior

Those files make the skill more actionable than a one-file prompt because they help you adapt output to the specific AI surface you care about.

Best-fit and misfit use cases

Best fit:

• “Why are competitors cited in ChatGPT and Perplexity, but not us?”
• “How should we structure pages for AI Overviews?”
• “Should we allow GPTBot or PerplexityBot?”
• “How do we rewrite SEO content so AI systems can quote it?”

Misfit:

• deep technical audits of site architecture
• schema markup implementation details
• backlink prospecting workflows
• purely editorial blog writing with no AI visibility goal

How to Use ai-seo skill

Install context for ai-seo

The repository evidence does not expose a built-in install command in SKILL.md, so the practical pattern is to add the parent skill repo and select ai-seo:

npx skills add https://github.com/coreyhaines31/marketingskills --skill ai-seo

After install, open the skill folder and read:

1. skills/ai-seo/SKILL.md
2. skills/ai-seo/references/platform-ranking-factors.md
3. skills/ai-seo/references/content-patterns.md
4. skills/ai-seo/evals/evals.json

That order gets you the workflow first, then ranking logic, then output patterns, then examples of what good execution should cover.

Read these files before your first prompt

If you only skim one file, read SKILL.md. If you want better output quality fast, add these:

• references/platform-ranking-factors.md to avoid treating Google AI Overviews and Perplexity as the same system
• references/content-patterns.md to turn vague advice into page blocks the model can actually draft
• evals/evals.json to see expected behaviors such as checking AI bot access, running a visibility audit, and prioritizing actions

This matters because ai-seo usage is better when the agent reasons from the repo’s framework instead of improvising from generic SEO knowledge.

Start with product marketing context if you have it

The skill explicitly tells the agent to check for .agents/product-marketing-context.md or older .claude/product-marketing-context.md before asking questions. That is important if your company already has positioning, ICP, differentiators, and proof points written down.

Without that context, the model may produce accurate but bland AI SEO advice. With it, the output can align content recommendations to the actual product narrative and buyer language.

Inputs ai-seo needs to work well

For a strong ai-seo guide, provide these inputs up front:

• target product or site
• 5 to 10 priority queries
• platforms you care about most
• whether you are already cited anywhere
• top competitors or commonly cited alternatives
• important URLs to audit
• whether AI bots are allowed in robots.txt
• content goals: definitions, comparisons, alternatives, how-to, statistics, FAQs

This skill is much stronger when it can evaluate a real visibility problem than when asked abstractly.

Turn a rough goal into a complete ai-seo prompt

Weak prompt:

“Help with AI SEO.”

Better prompt:

“Use the ai-seo skill to audit why our project management SaaS is not appearing in ChatGPT, Perplexity, and Google AI Overviews for queries like ‘best project management software for agencies’ and ‘Asana alternatives.’ Review our homepage, comparison pages, and product overview page. Assume we allow search crawlers but have not checked GPTBot or PerplexityBot. Give me: 1) a visibility diagnosis by platform, 2) highest-priority content fixes, 3) crawl/indexing checks, 4) page-block recommendations using answer-engine patterns, and 5) a 30-day action plan.”

That version gives the skill enough scope to apply its framework instead of falling back to generic tips.

Use the three-pillar workflow in practice

A practical ai-seo usage sequence looks like this:

1. Check presence
   Are you indexed and accessible on the platforms that matter?

2. Check structure
   Do your pages contain self-contained answers, comparisons, definitions, statistics, and FAQs that can be extracted cleanly?

3. Check authority
   Are claims sourced, specific, and written in a way that sounds citable rather than promotional?

4. Prioritize pages
   Start with high-intent commercial pages and comparison pages before broad thought-leadership content.

This is the main decision logic embedded in the skill and reinforced by the reference docs.

What content works best with ai-seo for SEO Content

ai-seo for SEO Content works best when the source material already has genuine substance. The support references point toward blocks that are easier for answer engines to lift:

• concise definition sections
• step-by-step blocks
• comparison tables
• FAQ blocks
• statistic-backed claims
• expert quote or evidence blocks
• self-contained answer paragraphs

If your page is mostly brand copy and feature claims, the skill can still help, but the output will often recommend structural rewrites before optimization.

Platform-specific guidance matters

One of the biggest reasons to use this skill is that it separates platform behavior. The references note that each system has different search backends and weighting. In practice that means:

• Google AI Overviews still leans on strong traditional SEO and trust signals
• Bing-connected ecosystems may behave differently from Google-based ones
• answer engines prefer passages they can extract and summarize cleanly

So do not ask for one universal AI SEO checklist if your target includes multiple platforms. Ask for a platform-by-platform plan.

Use evals as a quality bar

evals/evals.json shows what the skill expects a good answer to include. Useful examples from the eval evidence include:

• checking product marketing context first
• auditing visibility across major AI platforms
• checking content extractability
• checking AI bot access in robots.txt
• giving a prioritized action plan
• explaining the tradeoff of blocking AI crawlers

If your first result skips those points, ask the model to revise against the eval expectations.

Common adoption blockers before install

Most hesitation around ai-seo install is not technical. It is strategic:

• teams do not know which AI platforms matter most
• they have not tested whether they appear in answers today
• they cannot tell if the problem is indexing, structure, or authority
• they expect AI visibility to come from keyword tweaks alone

This skill helps most when you are willing to treat AI search as a sourcing and extraction problem, not just a ranking problem.

ai-seo skill FAQ

Is ai-seo only for big brands?

No. Smaller sites can benefit if they publish clear, specific, answerable content in a category where authority can be demonstrated. The skill is especially helpful for focused SaaS, services, and B2B sites that can create strong comparison, definition, and use-case pages.

Is ai-seo different from normal SEO prompts?

Yes. Generic SEO prompts usually optimize for rankings, keywords, and on-page basics. ai-seo is narrower and more useful when you need content that AI systems can discover, parse, trust, and cite. It also pushes you to check bot access and platform behavior, which standard prompts often miss.

Does ai-seo replace traditional SEO?

No. The skill assumes baseline SEO still matters, especially for platforms that rely on major web indexes. AI visibility usually builds on normal discoverability rather than replacing it.

Is this ai-seo skill beginner-friendly?

Mostly yes, if you already know your product and target queries. The concepts are practical, but beginners may need to slow down and verify basics like indexing, robots.txt, and current brand mentions in AI answers.

When should I not use ai-seo?

Do not start with ai-seo if your immediate task is:

• fixing technical crawl issues sitewide
• implementing structured data in detail
• doing a broad content calendar
• writing generic blog posts without a citation goal

It is most valuable when the question is specifically about AI answer visibility.

Should I block AI crawlers?

The repo evals suggest this is a core decision area, not a default yes or no. Blocking can reduce citation opportunities. Allowing can improve inclusion but may raise internal concerns about reuse. Use the skill to evaluate that tradeoff by content type, business model, and visibility goals.

How to Improve ai-seo skill

Give ai-seo evidence, not just objectives

The fastest way to improve ai-seo output is to provide live inputs:

• pages you want audited
• screenshots or notes from actual AI answers
• competitor examples that get cited
• current robots.txt
• target queries with search intent labels

A model can reason much better from “here are the pages and citations we’re losing” than from “help us show up more.”

Ask for output in decision-ready sections

Good prompt structure improves ai-seo usage quality. Ask for:

1. diagnosis
2. root causes
3. page-by-page fixes
4. content block rewrites
5. platform-specific notes
6. prioritized roadmap

That format prevents the model from spending too much space on background theory.

Improve extractability before polishing copy

A common failure mode is trying to make copy “more SEO-friendly” without making it easier to quote. Have the skill rewrite pages into extractable blocks first:

• one-sentence definitions
• plain-language summaries
• direct comparisons
• bullet criteria
• FAQ answers that stand alone outside page context

This often changes AI citation potential more than tone edits.

Add proof and specificity to strengthen authority

Another failure mode is unsupported claims. The references emphasize patterns like statistic citation blocks and evidence-backed claims for a reason: AI systems are more likely to use passages that sound attributable.

Better input:

“We reduce onboarding time by 37% based on 214 customer implementations.”

Worse input:

“We dramatically improve onboarding for modern teams.”

The first is more citable, more compressible, and easier to trust.

Iterate by platform, not with one generic revision

If the first draft is weak, do not just say “make it better.” Ask for iterations like:

• “Revise this for Google AI Overviews.”
• “Now adapt it for Perplexity-style citation behavior.”
• “Rewrite this comparison page to be easier for ChatGPT to quote.”

That forces the model to use the platform-ranking reference instead of flattening all answer engines into one.

Use the content patterns as building blocks

references/content-patterns.md is the most practical file for improving output quality. Ask the model to recast a page into named blocks from that file, such as:

• definition block
• step-by-step block
• comparison table block
• FAQ block
• evidence sandwich block
• self-contained answer block

This gives you content architecture, not just advice.

Validate against real AI visibility after changes

The best improvement loop for the ai-seo skill is operational:

• publish or revise page blocks
• test priority prompts on target AI platforms
• note whether your brand appears, is cited, or is omitted
• compare which passages are being used
• feed that back into the next prompt

Without that loop, you can improve the content on paper while learning little about actual answer-engine behavior.