Clear about the method

What your Visibility Score really measures

We check the technical signals visible on a public webpage, turn them into four readiness scores and explain what to improve. We also say clearly what those scores cannot prove.

The simple version

A technical website check—not a promise of AI recommendations

The scanner examines observable page structure: whether systems can find the page, extract clear answers, identify supporting evidence and discover safe interactions.

Those foundations matter, but reputation, relevance, the exact question and the AI service also affect whether a business appears in a live answer.

The scores do not prove that a live AI service currently cites, mentions or recommends your business.
What we check

Four technical-readiness dimensions

Each score groups repeatable signals found on the page you submit. The detailed report shows the individual findings.

S

Findability

Can crawlers discover, identify and interpret the page correctly?

  • Titles, descriptions and canonical URLs
  • Crawl and indexing signals
  • Mobile and performance foundations
  • Structured data and internal links
A

Answerability

Can a system extract a useful, direct answer from the content?

  • Clear headings and page purpose
  • Question-and-answer structure
  • Definitions and step-by-step content
  • Readable, specific explanations
C

Citation readiness

Are important claims supported and easy to attribute responsibly?

  • Named sources and evidence
  • Clear entities and authorship
  • Dates and freshness signals
  • Relevant topic coverage
R

Agent readiness

Are safe page actions understandable without hiding their effects?

  • Semantic controls and labels
  • Clear inputs and result structure
  • Visible user control
  • Read-only progressive enhancement
What the evidence supports

A useful structural diagnostic, with limits

The current versioned rule set passes a frozen local matrix covering positive, negative, ambiguous, unavailable, malformed and adversarial evidence for every rule.

An earlier score separated selected well-built and poorly-built sites in a small 40-site reference set, while a separate comparison with Perplexity citations found weak correlation. Those studies do not calibrate the revised score or prove live visibility.

Read the study and its limitations
44automated technical checks in the current scoring framework
40websites in the small structural reference set—not a universal benchmark
How to read the resultPass means the stated evidence was observed. Unknown means required evidence could not be verified and never earns points. Not applicable means the rule does not fairly apply. Use the report to brief the person who maintains your website, and measure live AI mentions separately.
From page to report

How a scan works

Timing varies with the website and scan route. No live AI recommendation is inferred from the structural score.

1

Read the public page

The scan inspects the public URL you submit and the signals available there.

2

Run the checks

Each signal is evaluated within one of the four technical-readiness dimensions.

3

Build the scores

Applicable findings are combined using the published versioned weights. The A-to-F letter is a conventional coverage band, not a calibrated probability, ranking or prediction. Unknown evidence earns no points; not-applicable rules are excluded.

4

Explain the next work

The report prioritises issues. Material changes should be reviewed before publication.

See what your website is ready for

Start with the free technical check. No signup or credit card is required.

Check my website