Entity Authority Score (EAS) Audit
A deep-dive, current-state diagnostic built on our proprietary systems model of AI search dynamics, with customized, mathematically verifiable recommendations to transform your AI visibility.
You Can't Manage What You Can't Measure
Your SEO agency hands you rankings. Your analytics platform shows you traffic. Your PR firm tracks share of voice. None of those tools tell you whether AI systems cite your organization as an authority when a buyer builds a shortlist. The EAS Audit closes that measurement gap.
The Entity Authority Score (EAS) is a 100-point diagnostic of how visible and credible your organization is to AI systems. It tells you exactly what is broken and what to fix first. Most organizations score between 35 and 55 on a first audit, firmly in the Absent or Doubt range. That is true even for companies with strong brand recognition and SEO performance, because AI authority infrastructure is distinct from web visibility infrastructure.
What the Score Measures
Four components, 25 points each. Together they answer the four questions AI systems ask before they cite you.
Identity Completeness
Can AI systems confirm who your organization is (name, history, category, structure) without hedging? Measured through structured data validity, authority database completeness, Knowledge Panel presence and persistent identifier chains.
Attribute Accuracy
Are your facts correct in machine-readable form? Founding date, leadership, location and defined activities, verified across registries.
Machine Readability
Can AI systems actually read your website? We test structured data validity, deployment breadth and W3C compliance.
Vocabulary Sovereignty
Do the terms that define your category trace back to your organization as the originator, in machine-readable attribution?
Together these components form the Authority Equation, and the score comes directly from Byrum's Dominance Inequality, the governing inequality for whether your organization shows up on AI-generated vendor shortlists.
What the Audit Covers
A current-state read of your entire AI authority posture, built from tests we run against the AI systems themselves.
01
Parametric recall baseline
With web browsing disabled, we submit five standardized category queries and count how often AI names you as a primary authority without hedging. This Web-Fetch-Disabled Recall Protocol shows what the model knows about you from training alone, and it is the foundation for everything else.
02
Perimeter-by-perimeter posture
We rate your identity, domain and vocabulary perimeters separately with the Per-Perimeter Posture Assessment, using the Entity Attribution Rate for each. A single composite score can hide a critical weakness: an identity attribution rate of 90% can sit beside a vocabulary rate of 0%.
03
Query coverage
The Variety Audit Protocol tests whether your structured data earns citations for every kind of query buyers use (category, comparative, problem-oriented and alternative formulations) and exposes the gaps competitors exploit.
04
Competitor map
We benchmark the entities you compete with and find the high-value positions nobody holds yet. See how to map and own competitors' gaps.
05
Score and tier
Your result maps to a stage on the LLM Ladder through the EAS Tiers: Absent (0-40), Emerging/Doubt (41-70), Cited (71-85) and Defended (86-100). Crossing 71 is where AI citation behavior shifts from hedged to confident.
What You Get
A prioritized roadmap
Recommendations sequenced by the Dependency Chain, because lower layers have to be complete before upper layers work, and classified by Durability so you invest in lasting gains instead of short-term lifts.
A revenue-at-risk view
We translate the score gap into a business number: annual revenue x 40% x (1 – CPQ/CPQ*). The 40% is a Big House Enterprise analytical estimate, so treat the result as directional. See the worked example.
A baseline to beat
AI retrains every three to six months, and entity signals decay between cycles unless maintained. Your audit becomes the baseline we re-measure against as the work lands.
The Math Behind the Score
The EAS is not a black box. It is defined in published, citable research: Entity Authority in AI Citation Systems introduces the EAS and Citation Probability at Query (CPQ), and Theorem 7, the Measurement Corollary grounds the EAS as a proxy for citation probability.
Those papers are candid about the limits. EAS precision is estimated at roughly plus or minus 25 percentage points of CPQ, and its validity weakens in high-competition categories, on non-neutral platforms, and when a team optimizes for the score instead of the underlying signal. Our recommendations are built around that, not around a vanity number. Explore the full program in The Ontological Dominance Papers.
Frequently Asked Questions
What does the EAS measure?
The composite health of your AI authority infrastructure across identity completeness, attribute accuracy, machine readability and vocabulary sovereignty, producing a single score that maps to an LLM Ladder stage and drives remediation priority.
Can the EAS be gamed?
Short-term, tactical investments can lift a score temporarily without building durable authority. The Durability Classification separates those from architectural investments that produce lasting improvement, and the score weights temporal and structural factors so it resists gaming.
How is this different from an SEO audit?
SEO targets search engine crawlers and human readers. An EAS Audit targets how AI systems recognize, remember and cite your organization, which depends on machine-readable identity, attribute accuracy and vocabulary ownership rather than rankings and backlinks.
How often should we re-measure?
We recommend quarterly measurement, in step with the pace at which signals decay between model training cycles, and a full Variety Audit at least annually or whenever a competitive displacement signal appears.
Explore the Rest of the Toolkit
Entity Engineering
The foundational offering. Optimize your machine-readable entity home and take control of your AI visibility.
Vocabulary Sovereignty
Tired of everyone stealing your proprietary language or ideas? Proprietary methods or offerings lost in a sea of noise? Vocabulary Sovereignty stakes a claim in the ground on truly novel territory for your brand or business to excel in AI search.
Psycholinguistics
A deep-dive branding exercise to transform your brand from generic and forgettable to truly unique and differentiated, all while staying true to your existing voice so the algorithms map your content to your entity.
Ready to Engineer Your Algorithmic Authority?
Start with a free 15-minute AI visibility snapshot. See exactly where you stand across ChatGPT, Claude, Perplexity and Google Knowledge Panels, and discover your entity recognition gaps before your competitors exploit them.