Readiness level: definition and formula

“Agent-ready” also needs the simulated agent to recommend the store, which it does only when the checks behind access, price, shipping and returns pass (on hotel and service sites, the booking or contact check instead).

Agent-ready: score ≥ 80, no critical blocker, recommended by the simulated agent · hard to read: score < 40 · otherwise: partly ready

Agent-ready also needs measurement coverage of at least 80%: the weight we measured divided by the weight that applies to the site, with every step of the agent’s path checked.

“Hard for agents to read” means a score under 40, whatever else is true. “Agent-ready” needs a score of 80 or more, no critical blocker, enough measurement coverage and, for a store, a recommendation from the simulated agent. Everything else is “partly ready”.

On a store, the simulated agent recommends it only when every measured check behind access, price, shipping and returns passed in full, with at least one measured check in each group; a single partly passed one there stops “agent-ready”. On a hotel site the booking entry check must pass instead, and on a service site the contact or quote form check.

The agent’s path has three steps: reach the site (robots.txt and security layer), read it (without JavaScript, and the product data) and buy (add to cart). If any step is “couldn’t check”, the level can’t be “agent-ready”, and an Insufficient measurement result stays “partly ready”.

• Fine differences: a score of 80 and a score one point lower can sit on different levels. Read the findings, not only the label.
• Whether assistants recommend the store in their answers: the level is about what our checks see on public pages. Answers are measured as the Mention rate.
• An AI model’s judgment: the simulated agent follows fixed rules over our check results, and no AI model decides the level.

• Clearing critical blockers and raising the score to 80 or more (Priority order).
• Getting the agent’s path measured: if a security layer stopped us on the product page, those steps stay unknown (Is bot protection open to agents?).
• On a store, completing partly passed checks on price, shipping and returns, for example Machine-readable shipping information and Machine-readable return policy, because the simulated agent needs them to pass in full.

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

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

Run the free test →All terms in the glossary →