Agent readiness score: definition and formula

Only checks we actually measured count.

score = round(100 × Σ (weight × credit) ÷ Σ weight), over measured checks only; then the critical-blocker cap

Our sample report (web-scraping.dev, October 8, 2026): measured checks carry 90 weight and earn 67.07; round(100 × 67.07 ÷ 90) = 75.

Source: Sample report · October 8, 2026

The score is the share of the measured weight a site earns, after any cap. A score of 80 or more can be “agent-ready” if the other conditions of the Readiness level hold; under 40 is “hard for agents to read”. One critical blocker caps it at 45 and two or more at 39; a cap only matters when the uncapped score is above it.

The five Category score values use the same formula inside each category, without the cap. Checks we couldn’t check leave both sides of the formula, so the same score can rest on more or less of your site; below 30 of 100 measured weight no score is shown (Insufficient measurement).

Compare scores from the same score model version (v2.6 now): when the model changes, a score can move with no change on the site.

• Whether AI assistants actually recommend you: the free test doesn’t ask them. That is the Mention rate, measured on paid plans.
• Anything beyond the pages we read: on a store we read one product page, so price and stock consistency are judged on that sample. Guest checkout can’t be seen from outside, so in the free test it is always “couldn’t check”.
• Why it changed between two runs with nothing changed on your side: if a security layer or a rate limit stops us in one run and not in the next, different checks are measured. A rate-limited check counts as “couldn’t check”, never as failed.
• An AI model’s opinion: no AI model takes part in the score, so the same pages give the same number.

• Fixing critical blockers first: while one is open, the score can’t go above 45. See Critical blocker.
• Then the failed and partly passed checks that lose the most points: the Priority order order lists them and the Potential score shows where the first three would take you. Each check has a page with what passes and how to fix it (all checks), and there are fix guides by shop platform.
• Letting our crawler read the site: checks that were “couldn’t check” become measured, which can move the score either way. How our crawler identifies itself.

See also: How the free test works · Sample report

Readiness score (free test): How 27 checks on your public pages become one number from 0 to 100.

• Check result and credit
• Couldn’t check
• Not applicable
• Insufficient measurement
• Critical blocker
• Readiness level
• Category score
• Priority order
• Potential score

Run the free test →All terms in the glossary →