UNIVERSITY · LESSON 5 · CHOOSING TOOLS · 4 MIN READ
How to read an agent readiness score
What goes into the number, and three things it cannot tell you.
Agent Readiness Compare editors · Spec status checked September 2026
Answer
A readiness score adds up the signals a scanner found, weighted by layer. To read one, look at the layer breakdown before the total, check which rules produced it and when, and treat it as a map of gaps rather than a measure of outcomes. A score does not show AI visibility, traffic or revenue.
On this page
1.What is inside the number?
A score is a weighted sum. ora.ai's methodology is a published example: four layers worth 20 points for discovery, 30 for access, 40 for usability and 10 for payments, with 16, 41, 56 and 6 checks. The total out of 100 maps to grade bands: A+ (95–100, Leading), A (86–94, Agent-Ready), B (70–85, Competitive), C (48–69, Needs Work), D (28–47, At Risk) and F (0–27, Unusable). is-agentic.com, where ora.ai runs the scans, uses tiers instead: essential checks worth 80 points, recommended capabilities worth 20, and up to 5 bonus points.
2.Why read the layers before the total?
Because a total can hide a weak layer. A site with strong discovery and access can post a respectable total while scoring low on usability, which is where an agent tries to operate, authenticate and complete a task. Compare each layer's points with its maximum, then read the failing checks inside the weakest layer. That is usually the order to fix things in.
3.Which checks count?
Scanners differ in what they include. ora.ai states that only checks it has verified agents rely on count, and its Deep Scan v2 (7 July 2026) separates Verified signals, backed by agent behaviour, from Emerging signals that are tracked but not yet scored. Cloudflare's Is It Agent Ready lists its five categories and the standards it checks but does not publish score bands. Two scanners can therefore grade the same site differently for defensible reasons, which is why we compare scanners on published criteria in the calculator.
4.Why does the date matter?
A score moves for one of two reasons: your site changed, or the rules changed. The standards moved several times in 2026, and scanners update their checks when they do. Record the date and the scanner's method with each score, so a later change can be traced to one cause or the other.
5.What can a score not tell you?
- Whether an agent completed your most valuable task. A task run answers that; see lesson 6.
- What happens behind sign-in. ora.ai states that its scans cannot evaluate login-gated content.
- How often you appear in AI answers, or whether readiness work changed sign-ups or revenue. Measure those with your own analytics.
Next: file checks and task runs.
Source: ora.ai methodology · ora.ai llms.txt · ora.ai, Deep Scan v2 · ora.ai, is-agentic.com with Vercel · isitagentready.com · Reviewed Sep 2026