Understand the offer
Can it identify what you provide, the relevant attributes and who provides it?
Agent Readiness
Can AI agents understand enough about your business to help customers act?
Agent Readiness uses published business information to assess whether an agent can understand your offer, establish suitability and support the right next step.
Four practical questions
What can an agent do with the information you provide?
Determine basic suitability
Can it find eligibility, exclusions, prerequisites and customer or market restrictions?
Identify the appropriate action
Can it find a stable quote, booking, application, demo, purchase or contact route?
Support safe initiation or handoff
Can it identify required inputs and the point where the user or a person should take control?
Two public concepts
Legibility and operability.
Agent Legibility
The offer and suitability conditions are clear enough for an agent to interpret. Useful evidence includes accessible facts, explicit exclusions and consistent product information.
Agent Operability
The next action can be located and safely initiated or handed off. Useful evidence includes clear input requirements, predictable entry points and appropriate confirmation.
Evidence-backed capability
Readable → Qualifiable → Routable → Initiable → Enabled
These cumulative descriptions state the highest capability supported by observed evidence. They are not badges or maturity levels.
Agent-readable
The offer can be interpreted.
Agent-qualifiable
Basic suitability can be established.
Agent-routable
The relevant next action can be identified.
Agent-initiable
Safe initiation or handoff is supported.
Agent-enabled
Comprehensive useful capability is evidenced across the relevant journey.
Where evidence does not establish a cumulative capability, the report says so. Agent-enabled does not claim successful transaction execution.
Capability > protocol
Useful capability earns its place.
No named technology is mandatory.
APIs, MCP, A2A and structured data matter where they create observable, useful capability. Their presence alone does not demonstrate readiness.
A clear, usable action route can be valuable without an API. A named protocol can exist while suitability information is still unclear.
Safe handoff can be the correct outcome.
- Autonomous checkout is not universally required.
- Professional and regulated services may need human confirmation.
- The relevant action might be a quote, application, booking or conversation.
- Do not infer Agent Readiness from what a model says about the business.