Framework
Five interconnected layers that earn AI citations — from earned authority to visibility measurement. Each layer compounds the one below it.
The foundation layer
Tier 1 media placements in publications AI engines trust and cite — Forbes, TechCrunch, Wall Street Journal.
Why it matters
Research shows 82–89% of AI-generated answers cite earned media over brand-owned content. AI engines trust third-party validation exponentially more than press releases or brand blogs. Without earned authority, the other four layers have nothing to amplify.
In practice
A fintech startup securing a TechCrunch feature. When someone asks ChatGPT "What are the best AI fintech tools?" — the TechCrunch article becomes a citation source. The brand enters the training corpus and recommendation pipeline.
How it connects: Feeds Entity Optimization (validating the brand entity across knowledge graphs) and Citation Architecture (creating quotable, AI-extractable content within those placements).
The identity layer
Structuring a brand's digital identity so AI systems can resolve, verify, and cite it consistently across platforms.
Why it matters
AI engines need to resolve your brand as a distinct entity before they can cite it. Inconsistent entity definitions, missing schema markup, or lack of knowledge graph presence means the AI sees your brand as fragmented or unverifiable.
In practice
A SaaS company with consistent Organization schema, verified knowledge panel, Wikidata entry, and unified entity references across Crunchbase, LinkedIn, and media mentions. AI systems can instantly verify and contextualize the brand.
How it connects: Turns Earned Authority signals into machine-readable identity proofs. Enables Citation Architecture by giving AI a clear, resolvable brand to cite.
The content layer
Content engineering for AI extraction — attribution magnets, quotable data, answer-first structure.
Why it matters
AI engines don't cite entire articles. They extract quotable fragments — statistics, one-sentence definitions, clear attributions. Without clean fragments to extract, even the best earned media results in zero citations.
In practice
A cybersecurity company's blog post starting with "74% of enterprise breaches originate from third-party vendor access." That sentence is an attribution magnet — clean, quotable, verifiable. AI engines extract it verbatim.
How it connects: Sits on Entity Optimization (so AI knows who to cite) and Earned Authority (so the source is trusted). Feeds into GEO/AEO by making content extraction-friendly.
The distribution layer
Generative Engine Optimization and Answer Engine Optimization — tactical optimization for AI search.
Why it matters
Even with earned authority, entity optimization, and citation-ready content, you still need distribution. GEO ensures content appears in ChatGPT, Claude, and Gemini. AEO ensures it appears in Perplexity and Google AI Overviews.
In practice
A B2B SaaS company ensuring comparison pages and research reports are indexed by AI crawlers with structured markup, optimized sitemaps, and semantic HTML. Perplexity cites them in answer results.
How it connects: Takes the foundation (Earned Authority), identity (Entity Optimization), and content structure (Citation Architecture) and ensures AI systems can find, parse, and surface it.
The feedback layer
Tracking citation frequency, recommendation rate, and brand share of voice across AI platforms.
Why it matters
Traditional PR metrics — impressions, AVE, media mentions — are vanity metrics in the AI era. What matters: How often does ChatGPT cite your brand? What's your recommendation rate on Perplexity? What's your share of voice across AI platforms?
In practice
A startup tracking 50 high-intent queries weekly. They discover a 12% citation rate vs. the category leader's 47% — and use that data to prioritize earned media in specific publications AI engines trust.
How it connects: Closes the loop. Reveals which layers are working, where citation gaps exist, and what to optimize next. Without measurement, MR is guesswork. With it, it's a compounding system.
Machine Relations is not linear. Each layer reinforces the others — creating an Algorithm Credibility Moat: a compounding advantage where each citation makes the next citation more likely. Like SEO domain authority, but for AI.
The Flywheel
Earned Authority validates → Entity Optimization resolves → Citation Architecture structures → GEO/AEO distributes → Measurement optimizes → repeat