Is the score based on AI or on fixed rules?

Fixed rules. The engine reads your pages and applies a versioned set of deterministic checks. The same site scored twice gives the same result, and we never send your pages to an AI model to be graded.

Does a high score mean I will be cited by ChatGPT?

No. The score measures how ready your site is to be found, trusted and quoted. It does not measure or guarantee actual mentions, citations or rankings in any AI product. No tool controls what an AI system says.

Why is Technical Accessibility only 5% of the score?

Because for most sites it is already fine and rarely the thing holding you back. It still matters as a gate: if a machine cannot fetch your pages at all, nothing else can be measured, and the report is marked provisional.

What if a signal does not apply to my type of site?

It is marked not applicable and excluded from the calculation rather than counted as zero. The engine also detects your site type (for example SaaS, local business or publisher) so location- or product-specific checks only apply where they make sense.

Reference

How scoring works

How Find Your AI Score turns your website into a 0 to 100 readiness score: the seven weighted components, the three readiness stages, and the deterministic rules behind every point.

Updated August 26, 2026

Find Your AI Score turns your website into a single number from 0 to 100. This page explains exactly how that number is built, so nothing about the score is a black box.

The overall score

The overall score is a weighted average of seven components. Each component is scored from 0 to 100 on its own, then combined using the weights below. The weights reflect how much each area tends to affect whether AI answer engines can use your site.

Component	Weight	What it measures
Entity Clarity	20%	How unambiguously the site says who the business is, what it does and where.
Offer Clarity	20%	Whether the products or services are stated plainly and in detail.
Prompt Coverage	20%	How well the pages answer the questions customers actually ask.
Sourceability	15%	Whether the content is specific, first-party and extractable as a source.
Structured Data	10%	Whether machine-readable schema states the facts accurately.
Evidence & Trust	10%	Identity, contact, policies and references that make the site accountable.
Technical Accessibility	5%	Whether a machine can actually fetch and read the pages.

The free preview shows all seven component scores and the overall number. The Premium audit adds the specific findings, the evidence behind each one, and a paste-ready fix.

The three readiness stages

Alongside the components, the report shows three diagnostic stages that map to what actually happens when an AI answer engine uses a page. These are derived from the components to help you see where in the pipeline you are weakest. They are shown separately and do not change the overall score.

Retrieval readiness. Can a system find and correctly identify the right page for a question?
Citation readiness. If the page is considered, is it trustworthy and specific enough to be used as a source?
Answer extractability. Can a clean, correct answer be lifted out of the page?

How each point is earned

Every point traces back to something we observed on your pages. The engine crawls up to twelve pages, extracts the visible text, structure and metadata, and applies a fixed set of rules. Three principles keep the score honest:

Deterministic. The same site scored twice gives the same result. There is no AI grading and no randomness.
Applicability-aware. A signal that does not apply to your type of site is excluded, not counted as zero, so you are not penalized for something irrelevant.
Versioned. The scoring model has a version, so a score you got last month stays reproducible, and changes to the method are tracked rather than silent.

Confidence is reported separately from the score. If the crawl was incomplete or a page could not be read, the report lowers confidence and can mark the result provisional, instead of pretending the site failed.

What the score does and does not tell you

The score is a diagnostic of your website’s structural readiness for retrieval, citation and answer extraction. It is not a promise of rankings, traffic, or actual mentions in ChatGPT, Gemini, Perplexity or any other AI product. What you can control is how ready your site is to be read and quoted, and that is exactly what we measure.

Frequently asked questions

Is the score based on AI or on fixed rules?: Fixed rules. The engine reads your pages and applies a versioned set of deterministic checks. The same site scored twice gives the same result, and we never send your pages to an AI model to be graded.
Does a high score mean I will be cited by ChatGPT?: No. The score measures how ready your site is to be found, trusted and quoted. It does not measure or guarantee actual mentions, citations or rankings in any AI product. No tool controls what an AI system says.
Why is Technical Accessibility only 5% of the score?: Because for most sites it is already fine and rarely the thing holding you back. It still matters as a gate: if a machine cannot fetch your pages at all, nothing else can be measured, and the report is marked provisional.
What if a signal does not apply to my type of site?: It is marked not applicable and excluded from the calculation rather than counted as zero. The engine also detects your site type (for example SaaS, local business or publisher) so location- or product-specific checks only apply where they make sense.