AI Readiness

AI readiness measures whether your page is legible to AI crawlers and answer engines. It is a checklist of 11 signals, each grounded in a specific captured artefact. Methodology version: 2026-08-14.

The 11 signals

The crawler-permission signals read your live robots.txt. The content signals read the HTML and JSON-LD that were actually fetched during the audit, not what you think is deployed.

The score

score = round(100 * passes / measured signals)

• Measured signals are pass, warning, and fail. A warning is measured but does not count as a pass: no partial credit.
• unavailable signals are excluded from the denominator. An unreachable robots.txt does not drag the score to zero.
• Bounds: 0..100. No signal measured at all: the score is null, shown as an unmeasured dash, not as Limited.

The grade

What this score is and is not

The checklist is experimental. It measures readiness signals: crawler permissions, structured data types, semantic landmarks, entity clarity, extractable HTML. It does not and cannot predict:

• Whether a real provider crawler visited your site.
• Whether your content gets cited in AI answers.
• Rankings, traffic, or revenue.

A spoofed user-agent request does not prove a real provider crawler was blocked or allowed, and robots permission does not prove a crawler came. If you want observed AI-answer behaviour rather than readiness signals, that requires external observation, which is reported separately with its own sample counts.

Improving the score

Each signal's finding links to the exact artefact that failed. Common fixes, in rough order of leverage:

Fixes are verified the same way everything else is: re-scan, compare signal by signal.