Capability readiness

Capability readiness belongs beside data readiness and model readiness. It asks whether the things an organization can do are legible and safe for an agent to do on someone's behalf.

Agents do not need your application; they need its capabilities. "We have APIs" is not the same answer. A callable action is not agent-ready until its meaning, permissions, constraints, side effects and recovery path are clear without relying on a human operator to fill the gaps.

A practical exercise

Take one workflow that matters and map the actions required to complete it, ignoring existing application boundaries. Then ask:

• Which actions are real business capabilities?
• Which are read-only, and which create side effects?
• Which need a person in the loop?
• Which can be reversed?
• Which require delegated identity rather than a shared service account?
• Which still depend on something only a human knows?

That exercise will tell you more about agent readiness than another demo. Tool design becomes product design, and the callable surface becomes part of the reasoning environment — which is also why capability is not authority, and why standards such as the Model Context Protocol matter.

Read more in The next AI interface may never be seen.