The Five AI Readiness Evidence Families ClickRadius Scores
ClickRadius AI Readiness is a 0–100 estimate of supported setup evidence observed for a submitted site. It contains exactly five weighted families: Off-site Authority, On-page GEO, Schema Engine, Entity / E-E-A-T, and Crawler Access. It does not contain SEO Health or AI Presence, and it does not estimate a site's probability of being cited.
The current score contract
The served model weights Off-site Authority at 28%, On-page GEO at 26%, Schema Engine at 18%, Entity / E-E-A-T at 18%, and Crawler Access at 10%. These are current starting estimates, not weights fitted to customer outcomes. Their precision must not be read as a measured causal share of citations, rankings, traffic, or revenue.
Coverage is part of the result. Missing evidence remains unknown rather than becoming zero or passing by default. When supported evidence does not clear the versioned minimum coverage threshold, AI Readiness remains unavailable instead of producing a confident number from a thin denominator.
The score reports what supported evidence was observed under a versioned model. It does not report what an external engine thought, saw, selected, or will do next.
— ClickRadius score boundary
Family 1: Off-site Authority (28%)
This family examines whether the places outside the submitted site that can describe the organization are set up and consistent: directory listings, the review base, syndication and social footprint, links from other sites, and the entity record. It concerns third-party and profile surfaces, not the site's own prose.
An observed listing or link is evidence that the surface exists. It does not prove that an AI provider crawled it, credited it, or used it in an answer. Off-site changes also require the third party's own authority and process; ClickRadius can guide and record that work but cannot silently control another publisher.
Family 2: On-page GEO (26%)
This family examines whether owned pages expose clear, specific, non-promotional passages that can be understood as answers. It includes passage structure, supported claims, quotations, statistics, and cited sources where appropriate.
The Princeton-led GEO study, presented at KDD 2024, reported visibility changes for several content interventions in its own benchmark. That research identifies properties worth inspecting; it does not fit the ClickRadius family weights or establish the effect of an edit on a particular customer, provider, prompt, or date. The benchmark began with the top five Google sources, randomly selected one source for optimization, and measured that source's relative visibility/impression in generated responses; it did not measure retrieval or citation-inclusion probability, and reported effects varied by domain.
Family 3: Schema Engine (18%)
This family examines supported structured data describing the business, services, articles, people, and other page types in machine-readable form. Valid syntax is not enough: markup must agree with the visible page and retained facts.
Schema can make a site's own declarations less ambiguous. It cannot compel an external engine to crawl, index, rank, recommend, quote, or cite the page.
Family 4: Entity / E-E-A-T (18%)
This family examines whether the available evidence describes a recognizable and credible entity: a complete entity record, identifiable authors with supported credentials, and consistency between owned claims and independent sources. An unavailable third-party observation stays unavailable; it is not inferred from the organization's own schema.
A family result is not proof that any external provider resolved or trusted the entity. It identifies the evidence ClickRadius could support at measurement time.
Family 5: Crawler Access (10%)
This family reports whether the site's robots policy declares access for relevant AI crawlers and whether the site publishes an llms.txt. The distinction between policy and behavior is essential: permission does not prove that a crawler arrived, rendered JavaScript, indexed content, or cited a page.
Two measurements that remain outside AI Readiness
SEO Health separately reports conventional search hygiene such as metadata, technical soundness, security, performance, and supported coverage. It is not blended into AI Readiness.
AI Presence separately summarizes monthly sampled answers from ChatGPT, Gemini, Perplexity, Claude, and Grok. Microsoft Copilot is not currently supported. A defensible report retains the tested prompt set, successful-response denominator, mentions, citations, and unknown or failed observations for each engine. It is a sample, not continuous monitoring or a census of every answer users received.
AI Presence can establish what the configured sample observed. It cannot establish why an answer changed or whether a prior revision caused that observation.
How to read a score without overclaiming
- Check identity and version. Confirm the submitted site, scan time, score version, and evidence age before comparing results.
- Check coverage. Distinguish measured zero, unavailable evidence, not-applicable evidence, and a withheld composite.
- Inspect all five families. The composite can hide which supported observations changed.
- Keep adjacent measures separate. SEO Health and monthly AI Presence answer different questions.
- Use findings to prepare work, not predict outcomes. External engines retain control over crawling, indexing, retrieval, ranking, recommendation, and citation.
There is no universal score band that proves a site is citable, and a change in the number is not ROI. A comparison is meaningful only when the target, score version, and evidence basis are sufficiently comparable.
From a finding to a verifiable site effect
A finding may support a proposed revision. It does not authorize that revision. ClickRadius keeps the stages distinct: the exact revision is prepared, current authority is confirmed, the connection is proven capable of that effect, a provider receipt is retained, and an independent public-origin reread is performed where supported. Provider acceptance is not a public effect, and a correction or withdrawal path must remain available.
Frequently asked questions
What are the five evidence families ClickRadius scores?
AI Readiness comprises Off-site Authority (28%), On-page GEO (26%), Schema Engine (18%), Entity / E-E-A-T (18%), and Crawler Access (10%). These are current starting estimates, not weights fitted to customer outcomes. SEO Health and AI Presence are separate measurements.
Does AI Readiness measure whether an engine will cite a site?
No. It estimates supported setup evidence observed for the submitted site. External engines independently decide what to crawl, index, retrieve, rank, recommend, and cite. ClickRadius reports monthly sampled AI Presence separately rather than treating a readiness result as citation evidence.
What happens when evidence is missing?
Missing evidence remains unknown rather than becoming zero or passing by default. When measured coverage does not clear the versioned minimum, AI Readiness remains unavailable instead of producing a confident score from a thin denominator.
Submit a URL for a free AI Readiness assessment, or review ClickRadius plans for governed correction preparation and separate monthly AI Presence samples. The assessment measures supported setup evidence; it does not contact the five engines or grade citability.