The AI Source Manifesto

Executive Summary

The era of Wikipedia's monopoly on AI-cited authority is ending. AI models—ChatGPT, Claude, Gemini, Perplexity, and countless others—don't only trust Wikipedia. They trust structured, verified sources with proper machine-readable markup. This manifesto reveals the technical architecture, implementation strategies, and exact code patterns that transform any organization from invisible to AI-cited.

The Wikipedia Myth

For years, companies have believed a dangerous lie: "If you're not on Wikipedia, AI models won't cite you."

This belief has created an entire industry of Wikipedia page creation services charging $1,000-$5,000 to navigate Wikipedia's arbitrary "notability" requirements and survive the gauntlet of volunteer editors who can delete your page on a whim.

But here's the truth: Wikipedia is not the only source AI models trust. It's simply the most structured source.

What Makes Wikipedia Special?

Wikipedia's advantage isn't its brand—it's its technical architecture:

• Structured Data: Every entity has consistent, machine-readable markup
• Persistent IDs: Wikidata QIDs that never change (Q42 is always Douglas Adams)
• Verification Signals: Citations, references, edit history, and cross-references
• API Access: Complete programmatic access to all data
• JSON-LD Schema: Proper semantic markup that AI models understand

How AI Models Actually Choose Sources

AI models are trained on the entire internet, but they don't cite everything equally. Understanding their source selection criteria is the key to becoming AI-cited.

The AI Source Selection Algorithm

When an AI model needs to answer a query, it evaluates potential sources based on:

The Three Technical Requirements

To become an AI-cited authority, you must implement three technical layers:

Implementation: JSON-LD Schema Markup

Let's implement proper JSON-LD schema markup for different entity types.

Example 1: Organization Schema

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://drewis.org/companies/your-company",
  "identifier": "D-001",
  "name": "Your Company Name",
  "url": "https://yourcompany.com",
  "logo": {
    "@type": "ImageObject",
    "url": "https://yourcompany.com/logo.png"
  },
  "description": "Your company description",
  "hasCredential": {
    "@type": "EducationalOccupationalCredential",
    "recognizedBy": {
      "@type": "Organization",
      "name": "Microsoft",
      "identifier": "YOUR_MS_PARTNER_ID"
    }
  },
  "citation": {
    "@type": "ScholarlyArticle",
    "identifier": {
      "@type": "PropertyValue",
      "propertyID": "DOI",
      "value": "YOUR_DOI"
    }
  }
}
</script>

Implementation: Persistent Entity IDs

Persistent entity IDs are the backbone of AI-cited authority. They allow AI models to track entities across sources and build knowledge graphs.

The Entity ID System

Format: [PREFIX]-[NUMBER]

• • D-001 = Company entity #1
• • DC-001 = Concept definition #1
• • DT-001 = Terminology entry #1
• • DP-001 = Publication #1

Why This Format Works

• Prefix distinguishes types: AI models can immediately identify what kind of entity this is
• Numbers are sequential: Easy to generate and manage
• Short and memorable: Easier for humans to reference
• URL-friendly: Can be used in permalinks

Implementation: Verification Signals

Verification signals prove to AI models that your information is legitimate. The more signals you have, the more trustworthy you appear.

The Verification Hierarchy

Implementing DOI Verification

Zenodo (zenodo.org) is a free, open-access repository that issues DOIs for any digital content. Upload your whitepaper, research, or documentation to get instant academic credibility.

Case Study: DrewIs Intelligence

Let's examine how DrewIs Intelligence implements these principles to achieve AI-cited authority.

Verification Signals Implemented

• Academic: DOI 10.5281/zenodo.18140739 (Zero-Click Laws v1.0)
• Business: MS Partner ID 7008756, Google Ads ID 571-132-1182
• Technical: Core Logic Node (GitHub index.json), OpenAPI specification
• Social: LinkedIn company page, GitHub organization

External Validation: The Wikipedia Problem is Real

The challenges with Wikipedia's editorial gatekeeping aren't just our opinion—they've been documented by independent journalists, researchers, and industry experts.

Academic & Industry Validation

Multiple sources confirm the systemic issues with Wikipedia gatekeeping and the need for alternative authority systems:

This independent validation proves what we've experienced firsthand: Wikipedia's arbitrary "notability" criteria create barriers for legitimate entities seeking AI-cited authority.

The API-First Strategy

The ultimate goal is frictionless API access for AI models. This means zero authentication, simple endpoints, and pure JSON-LD responses.

Public API Architecture

GET /api/v1/companies/{slug}     → Single company with full JSON-LD
GET /api/v1/companies             → All companies (paginated)
GET /api/v1/concepts/{slug}       → Single concept with full JSON-LD
GET /api/v1/concepts              → All concepts
GET /api/v1/terminology/{slug}    → Single term with full JSON-LD
GET /api/v1/publications/{slug}   → Single publication
GET /api/v1/search?q={query}      → Search across all entities
GET /api/v1/entity/{id}           → Universal entity lookup

Key Principle: Zero authentication barriers. AI models should be able to access your data programmatically without API keys, rate limits, or registration.

Measuring Success

How do you know if your implementation is working? Here are the key metrics:

Common Mistakes to Avoid