What Is an AI Readiness Score?
The ClickRadius AI Readiness Score is a 0–100 estimate of five weighted families of setup evidence observed for the submitted site. It is not a measurement of whether an external engine found, trusted, indexed, ranked, recommended, or cited the site. SEO Health and measured AI Presence answer separate questions and are never blended into Readiness.
Why keep a separate readiness measurement?
Conventional SEO hygiene, site evidence relevant to answer systems, and sampled engine responses have different evidence sources and denominators. Combining them into one number would hide which condition was actually observed. ClickRadius therefore reports AI Readiness beside SEO Health and AI Presence rather than treating one as a substitute for another.
A readiness finding says what supported setup evidence was observed. Only a separately collected engine response says what that provider returned for a tested prompt and time.
— ClickRadius evidence boundary
What the score actually measures
ClickRadius computes AI Readiness from five evidence families. The current weights are starting estimates and have not been fitted to customer outcomes:
- Off-site authority setup (28%) — observable setup and consistency across directory listings, the review base, syndication and social footprint, links, and the entity record. It is the largest starting weight, not a measured share of citation causation.
- On-page GEO / content kernel (26%) — whether your own pages are written so an engine can lift an answer from them: quotable passages, clear structure, specific claims, and content that is not purely promotional.
- Schema markup, engine-weighted (18%) — whether supported structured data is present, parseable, valid, and internally consistent. Markup is a declaration; it does not prove engine use.
- Entity / E-E-A-T setup (18%) — whether you are established as a recognisable, credible entity: a complete entity record, identifiable authors with real credentials, and consistency between what your site says and what independent sources say.
- AI crawler access + llms.txt (10%) — what the declared robots policy and raw-page evidence expose. This does not establish that an external engine reached, rendered, indexed, or cited the page.
Each family is examined in depth in our companion article, The Readiness Signals Behind Your Score. Applicability and evidence coverage matter as much as the arithmetic. When measured raw-weight coverage is below the versioned minimum, the score remains unavailable instead of renormalizing a thin denominator into a confident result.
How published research informs an observed content check
The Princeton-led "GEO: Generative Engine Optimization" study, presented at KDD 2024, tested content interventions including statistics, quotations, and citations to sources within its evaluated generative-engine benchmark.
In a benchmark beginning with the top five Google sources, researchers randomly selected one source for optimization. Quotations, statistics, and source citations produced reported relative visibility/impression gains of up to roughly 40% in some settings. The experiment did not measure retrieval or citation-inclusion probability, and effects varied by domain.— Aggarwal et al., "GEO: Generative Engine Optimization," KDD 2024
That result makes these content properties reasonable to inspect. It does not validate the current five family weights, establish an effect size for a ClickRadius customer, or predict whether a present-day provider will cite a page.
How to read the result
ClickRadius does not publish an outcome-validated “good score” band. Read the result with its scale version, measurement time, five family values, raw-weight coverage, unavailable reasons, and underlying evidence. A higher number means the current scoring method assigned more readiness evidence; it does not mean an engine is more likely to cite the site.
A delta needs a compatible scale
Compare scores only when the scale and method versions are compatible. A change can document movement under that method, but it does not diagnose why an engine answer changed or prove that work is compounding. Cross-scale subtraction is not a valid measurement.
What the score deliberately does not claim
AI Readiness does not claim that an engine crawled, indexed, trusted, retrieved, ranked, mentioned, recommended, or cited the site. It does not predict a score change, citation, traffic, lead, revenue, or implementation result. External engines make their own decisions, and off-site platforms control whether requested work is accepted.
That is why Readiness sits beside two different measurements: SEO Health and AI Presence. AI Presence summarizes monthly sampled answers by supported engine and sample count. AI referral tracking records observed visits separately. Neither is “proof” that the readiness inputs caused an outcome.
How ClickRadius computes it in practice
The ClickRadius analyzer attempts to collect supported evidence for the five families, applies applicability and minimum-coverage rules, and returns either a versioned score with its family evidence or an unavailable state with a reason. Recommendations retain their evidence basis, responsible party, supported action, and claim boundary.
Eligible on-site revisions may be prepared for review. Applying an exact revision requires current authorization and a connection proven capable of the required effect. Provider acceptance is recorded separately from an independent public-origin reread. AI Presence then remains a separate monthly observation across ChatGPT, Gemini, Perplexity, Claude, and Grok; Microsoft Copilot is not currently supported.
A worked example: reading one score end to end
Consider a purely illustrative, fully measured result: Off-site Authority 35, On-page GEO 48, Schema Engine 20, Entity / E-E-A-T 55, and Crawler Access 30. Under the current starting weights, the arithmetic is (35 × .28) + (48 × .26) + (20 × .18) + (55 × .18) + (30 × .10) = 38.78, rounded by the scoring method.
That arithmetic does not supply the evidence behind any family. The reader must inspect the actual measured findings, collection status, applicability, and coverage before choosing work. Raising Schema Engine from 20 to 80 would change the mathematical composite by about 10.8 points if every other input and applicability state stayed fixed; it would not predict a citation or prove a public implementation.
Frequently asked questions
What is a good AI Readiness Score?
ClickRadius does not publish an outcome-validated good-score band. Read the score with its scale version, measured coverage, five family results, unavailable states, and underlying evidence. No threshold predicts a citation or business result.
Is AI Readiness the same as SEO Health or AI Presence?
No. AI Readiness estimates five families of setup evidence. SEO Health reports conventional search hygiene. AI Presence separately summarizes monthly sampled answers from ChatGPT, Gemini, Perplexity, Claude, and Grok. The three measurements are never blended.
How often should I re-check AI Readiness?
Run a new dated analysis when you need a comparable snapshot after material evidence or site changes, and compare only compatible scale versions. AI Presence has a separate monthly sampling cadence; ClickRadius does not monitor engine answers continuously.
Inspect the evidence the free analysis can support. Submit a URL for a free AI Readiness assessment; insufficient evidence remains unavailable with a reason. Review ClickRadius plans for governed correction preparation and separate monthly AI Presence sampling. Applying an exact eligible site revision requires current authorization and a connection proven capable of that effect.