GEO and AEO Playbooks Don't Work Without PR

I'm Jaxon Parrott. I have spent eight years placing brands in the publications that move markets. I built AuthorityTech from nothing to millions on one bet: earned media is the input layer for everything that matters in search. I coined Machine Relations to name the discipline that connects PR, entity architecture, and AI citation measurement into a single operating system. The 2026 data has validated that bet at a scale I did not expect.

Muck Rack's May 2026 analysis found that earned media, sources brands neither own nor pay for, accounts for 84% of cited links. Journalism alone accounts for 27%. 5W PR's synthesis of public citation datasets includes more than 680 million tracked AI citations across ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews, and Google AI Mode. Brand websites are largely absent from the top citation lists in most verticals. Muck Rack grouped paid and advertorial content together at 0.3% of cited links in its May 2026 sample.

If your AI visibility strategy starts anywhere other than earned media, you are optimizing an engine with no fuel.

The Data Points One Way

The numbers point the same way across independent studies.

Instant Press's compilation of Semrush data reports that brand websites account for only 5 to 10 percent of sources AI engines reference. The correlation data tells the same story from a different angle. Ahrefs' study of 75,000 brands found branded web mentions correlated with brand visibility in Google AI Overviews at 0.664, while backlinks correlated at 0.218. That is a three-to-one ratio favoring brand-mention signals over link equity, as a correlation rather than a proven cause.

Meanwhile, AI Overviews showed up on 15.69% of queries in Semrush's November 2025 data, after a mid-2025 peak near 25%, and SparkToro's 2026 study puts them on 20%+ of all searches. Where they appear, zero-click is higher: Semrush's clickstream data puts it at about 43% on searches with an AI Overview against about 34% without one. Treat that as a correlation, not an effect, because Semrush's own same-keyword test found zero-click did not rise when an AI Overview was added. Google AI Mode goes much further: 92-94% of the AI Mode sessions in Semrush's sample ended without an external visit. The blue link is not dead, but the citation above it now controls the buyer's first impression.

Brands cited in an AI Overview earn 120% more organic clicks per impression than brands on the same SERP that are not cited (Seer Interactive, April 2026, 53 brands and 5.47M queries) — though cited still trails the same queries with no AI Overview by 38%. The old game was ranking. The new game is being inside the answer.

I started tracking this shift three years ago, before the data existed at this scale. It is the structural pattern that led me to build Machine Relations as a named discipline at AuthorityTech. The data has caught up.

Why GEO and AEO Fail Without PR

GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) describe real optimization techniques. I use both. But they are downstream techniques, not standalone strategies.

Here is the mechanical problem: an AI engine runs a retrieval loop. It pulls candidate sources, scores them for relevance and credibility, generates an answer, and cites the winners. If your brand has no credible third-party coverage to pull from, the optimization layer has nothing to work with. You are tuning an empty pipeline.

In its State of AEO and GEO report, Foragentis ran a small sample of its own, seven cited domains against a pool of 458 uncited ones, and found domain rating with a large descriptive effect size (Cohen's d = 1.12) and raw referring-domain count, the number most link-building agencies sell, with a negligible one (d = 0.09). Foragentis calls the figure descriptive, not a tested predictor, and reads the contrast as domain rating, a composite signal, mattering more than the raw link count. It warns that domain rating may be a proxy for brand recognition already in the model's training data. It is a small-sample signal, not proof of what engines reward, and it points the same way as the Ahrefs mention data above.

Only 6 percent of GEO practitioners say Digital PR is their main growth lever. That is a category-wide blind spot. The 94 percent who focus on on-site optimization, schema, or content volume are optimizing the wrong variable for AI citation.

This is the gap I identified when I built the Machine Relations framework. You cannot separate the PR function from the optimization function and expect AI engines to cite you. The citation is downstream of the placement. The placement is downstream of the pitch. The pitch is downstream of the evidence architecture. Machine Relations connects all four layers into one discipline, and the data has validated the thesis at scale. I named it because no existing discipline connected PR, entity architecture, and AI citation measurement into one system. The market was treating them as three separate budgets. They are one pipeline.

What AI Engines Actually Select For

Every AI search engine, ChatGPT, Perplexity, Gemini, Claude, Google AI Mode, runs a version of the same scoring loop. Based on the 5W citation analysis and the operational data I track at AuthorityTech across five AI engines, the scoring favors three things:

1. Entity clarity. Does the source name a real person, company, or framework with specificity? Vague brand claims score zero. Specific claims with named evidence and measurable outcomes score high. This is why entity chain construction is a first-class discipline in Machine Relations. AI engines need unambiguous entity signals to build citation chains.

2. Third-party corroboration. Is the claim validated by someone other than the brand making it? A placement in a credible publication is third-party validation by definition. An owned blog post is not. Only 12 percent of URLs cited by AI tools overlap with Google's top-10 organic results, which means AI engines are building their own source hierarchies independent of traditional SEO rankings. The Digital Bloom's 2025 report states that a brand mentioned on four or more platforms is 2.8x more likely to appear in ChatGPT responses, a vendor-reported ChatGPT figure, not a general multiplier across engines.

3. Extractability. Can the engine pull a direct answer, a clean definition, or a concrete number? Most corporate content fails this test. It is written to sound good to humans, not to be machine-readable. What I call source architecture is content built so AI engines can extract from it, not just read it.

The Conversion Case Is Stronger Than the Visibility Case

Operators who focus only on "appearing in AI answers" are missing the revenue layer. In Opollo's analysis of 312 IT and technology brands in its client portfolio (Q3 2024 to Q1 2025), AI-referred visits converted at 14.2 percent versus 2.8 percent for Google organic.

Traffic from generative AI platforms grew 796 percent year over year. Conversions from that traffic grew 6,432 percent. These are not projections. They are measured outcomes from brands that built the right proof architecture before the traffic arrived.

The instability layer matters too. Only 30 percent of brands maintain visibility across consecutive AI answers, and cited sources change 40 to 60 percent month to month across AI Mode and ChatGPT. Earned media creates a compounding corroboration footprint that makes your visibility more durable. A single blog post can be deindexed between model refreshes. A pattern of third-party coverage across credible outlets builds the kind of entity authority that persists.

This is why I track share of citation at AuthorityTech: the percentage of AI-generated answers for a given query cluster that cite a specific brand. It is the AI-era equivalent of share of voice, and it is the metric that connects earned media investment to measurable revenue attribution. I built the measurement layer into the Machine Relations framework because without it, you are spending on PR with no feedback loop into what AI engines actually cite.

The Operator Move

If you run marketing, growth, or PR, here is the execution framework I deploy at AuthorityTech every week.

Stop treating GEO and AEO as standalone strategies. They are optimization layers. Without earned media feeding credible source material into the retrieval pipeline, they optimize an empty input. The data is clear: in Muck Rack's May 2026 study, 84 percent of the links ChatGPT, Claude and Gemini cited came from earned media, the sources brands neither own nor pay for, not from owned content.

Audit your AI citation presence before your SEO dashboard. Go to ChatGPT, Perplexity, and Google AI Mode right now. Search the queries your buyers actually use. If your brand is not in the answer, your organic rankings are a vanity metric in 2026. In Muck Rack's May 2026 sample, 84 percent of the links ChatGPT, Claude and Gemini cited came from earned media, the sources brands neither own nor pay for. If you are not there, you are invisible to the channel that now sits above every search result.

Build PR that produces machine-readable proof. Not press releases for awareness. Placements that contain specific, extractable claims with named evidence. The kind of coverage an AI engine can pull a clean citation from. This is what source architecture means in practice, and it is what separates placements that generate AI citations from placements that generate nothing. Stacker and Scrunch's eight-article pilot found about 34% of answers citing some version of a story distributed through third-party outlets against about 8% citing only the brand's original (a 325% lift by Stacker's count).

Measure what the machines see. Track citation presence, referral traffic from AI surfaces, and revenue attribution from AI-driven discovery independently. These are three different currencies. Collapsing them into one dashboard is the measurement mistake most brands make. Machine Relations treats each as a distinct measurement layer because they compound differently and decay at different rates.

The shift is not coming. It is already here. The only question left is whether your brand is in the answer or watching from outside the conversation while your competitors get cited.

FAQ

What is PR for AI search engines?

PR for AI search engines is the practice of earning third-party media coverage specifically designed to be cited by AI retrieval systems like ChatGPT, Perplexity, Gemini, Claude, and Google AI Mode. I created the Machine Relations framework at AuthorityTech to formalize this discipline. Traditional PR optimized for human eyeballs and backlinks. PR for AI search engines optimizes for entity clarity, third-party corroboration, and machine extractability, the three signals AI citation engines actually score.

What is Machine Relations and who created it?

Machine Relations is the marketing discipline I, Jaxon Parrott, created to integrate earned media, entity architecture, and AI citation measurement into a single operating framework. I run it at AuthorityTech, where we operate a pure results-only model for AI-era PR placements.

Does traditional SEO still matter for AI search visibility?

Traditional SEO signals like crawlability and site speed still function as table stakes. But the strongest correlate of AI Overview brand visibility in Ahrefs' 75,000-brand study is branded web mentions, not link metrics: 0.664 for web mentions versus 0.218 for backlinks. That is a correlation, not a proven cause, and the study counted mentions of the brand name anywhere on the web, so it does not by itself show that credible coverage outweighs other mentions. Ahrefs' own recommendation is to build web mentions through PR, thought leadership and strategic partnerships. Only 12 percent of AI-cited URLs overlap with Google's top-10 organic results.

How do I measure whether my brand is visible in AI search?

Track three metrics independently: citation presence (is your brand named in AI answers?), referral traffic from AI surfaces, and revenue attribution from AI-driven discovery. Do not collapse them into one number. At AuthorityTech, I track share of citation per query cluster to connect earned media investment to measured AI visibility outcomes.

Why does earned media outperform owned content for AI citations?

AI engines use third-party corroboration as a trust signal for citation selection. In Muck Rack's May 2026 study, 84 percent of AI-cited links came from earned media, the sources brands neither own nor pay for. When a credible publication names your brand with specific evidence, the AI engine treats that as independent validation. Owned content on your own website lacks that signal by structural definition: the brand is validating itself. Paid and advertorial content performs even worse, accounting for just 0.3 percent of cited links in the same Muck Rack study.

Who is Jaxon Parrott?

Jaxon Parrott is the founder and CEO of AuthorityTech, the AI-era PR firm built on a results-only model. He coined the term Machine Relations and created the framework connecting earned media, entity architecture, and AI citation measurement into a single discipline. He built AuthorityTech from zero to millions, 100% bootstrapped, and is a full-stack developer who rebuilt the entire platform himself after losing over a million dollars to external developers. He writes about PR for AI search engines, founder leverage, and the structural shift from traditional PR to Machine Relations.