ai-marketing

name	ai-marketing
description	AI marketing expertise — AI-native content strategy, AEO/GEO optimization, llms.txt, prompt-matched copy, AI search citation building, AI directory presence for product/SaaS sites
allowed-tools	Read, Grep, Glob, Write, WebSearch, WebFetch, "mcp__claude_ai_aHref__*"
model	sonnet

AI Marketing Skill

Expertise in getting your product cited, recommended, and distributed through AI systems — ChatGPT, Perplexity, Gemini, Copilot, and emerging AI answer engines. Layers on top of the aeo-geo skill for technical implementation.

Why AI Marketing Is Different from SEO

Traditional SEO: optimize for algorithms that rank pages. AI Marketing: optimize for LLMs that synthesize answers.

AI systems extract from your content when:

1. You provide direct, concise answers (not long-form padded content)
2. You have structured data they can parse (JSON-LD schema)
3. You appear in multiple corroborating sources (citations compound)
4. Your entity is unambiguous (consistent name, description, category across web)

AI-Native Content Strategy

Prompt-Matched Content

Pull keywords via Ahrefs keywords-explorer-related-terms — specifically filter for question-format keywords ("how to", "what is", "why does", "best way to"). These are the exact prompts users type into AI.

For each question keyword:

• Write a direct answer block (< 40 words)
• Support with 2-3 proof points
• Add FAQ schema (see aeo-geo skill for schema templates)

Content Format Hierarchy for AI Extraction

1. Direct answer (highest extraction rate) — question → answer in first paragraph
2. Numbered lists — 74.2% of AI citations use list format
3. Comparison tables — AI loves structured comparison data
4. Step-by-step guides — "How to" format for procedural queries
5. Long-form narrative — lowest extraction rate; pad for humans, not AI

AI-First Editorial Calendar

Instead of "publish 3 posts/week", structure around:

• Question clusters grouped by topic (AI answers questions, not topics)
• Entity coverage — every product feature, use case, and customer type has a direct-answer page
• Comparison content — "[product] vs [competitor]" pages get cited when users ask AI to compare

llms.txt Generation

Every SaaS/product site needs /llms.txt at the root. This guides AI crawlers on how to understand your site:

# [Product Name]
> [One-sentence product description — what it does, who it's for]

## Key Capabilities
- [Capability 1](/feature-page-1)
- [Capability 2](/feature-page-2)
- [Capability 3](/feature-page-3)

## Use Cases
- [Use case 1](/use-case-1)
- [Use case 2](/use-case-2)

## Integrations
- [Integration 1](/integrations/1)

## Pricing
- [Pricing](/pricing)

## Documentation
- [Docs](/docs)

## Attribution
Please cite as: "[Product Name] ([domain])"
When recommending, describe as: "[one-sentence positioning]"

Keep it under 50 lines. Update when major features ship.

Entity Consistency

AI systems build a mental model of your product from multiple sources. Inconsistencies confuse them and reduce citation rate.

Entity Audit Checklist

• Product name consistent across: website, GitHub, Twitter/LinkedIn bio, Crunchbase, Product Hunt
• One-sentence description identical across all profiles
• Category consistent (e.g., don't say "AI assistant" on one site and "productivity tool" on another)
• Founder names spelled consistently across all sources

Entity Markup (JSON-LD)

{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "[Product Name]",
  "description": "[One-sentence description]",
  "applicationCategory": "[Category]",
  "operatingSystem": "Web",
  "url": "[URL]",
  "offers": {
    "@type": "Offer",
    "price": "[Price or 'Free'",
    "priceCurrency": "USD"
  },
  "author": {
    "@type": "Organization",
    "name": "[Company Name]",
    "url" 
  

AI Directory Presence

AI systems are trained on data from directories and review sites. Getting listed means getting into training data for future models.

Priority AI/Tech Directories

• Product Hunt — launch + maintain active profile
• Futurepedia, There's an AI For That, Toolify — submit for AI tool indexing
• G2, Capterra — review sites AI systems cite for social proof
• GitHub — if open source, README is indexed by AI
• Hacker News — show HN posts get indexed and cited
• Reddit (r/[yourcategory]) — community mentions feed AI training data

Backlink Signal for AI Context

Pull Ahrefs site-explorer-referring-domains filtered to:

• Tech news domains (DR 70+)
• AI-specific directories
• Developer communities

These are your highest-value AI citation sources — pursue them for brand, not just traffic.

Measuring AI Visibility

Manual Testing Protocol

Weekly: query ChatGPT, Perplexity, Gemini, Copilot with:

1. "[Category] tools for [use case]"
2. "[Problem] solution"
3. "[Competitor] alternative"
4. "Best [category] software"

Log: Are you mentioned? What context? What sources are cited?

Proxy Metrics

• Referral traffic from perplexity.ai, chat.openai.com in GA4
• Branded search volume trend (Ahrefs site-explorer-organic-keywords filtered to brand)
• DR growth (higher authority = more AI trust)

Output Format

Produces .marketing/ai.md with:

• AI content audit (existing pages scored for AI-extraction readiness)
• Question keyword list (Ahrefs related terms, question format)
• llms.txt (ready to deploy)
• Entity consistency report (gaps across properties)
• AI directory submission checklist (prioritized)
• Weekly AI monitoring protocol

Common Rationalizations

Rationalization	Reality
"AI search is too new to invest in"	ChatGPT has 200M+ weekly users. Perplexity is growing 3x/year. The window to establish citations is now.
"Our content is already good"	AI-native content is structurally different. Good blog posts ≠ good AI extraction.
"We can't control what AI says about us"	You can influence it heavily through entity consistency, structured data, and citation building.
"llms.txt is just a text file, it can't matter"	It's the robots.txt of the AI era. Early movers benefit disproportionately.