Independent AI capability registry

AI capabilities are changing inside the software organisations already use.

An independent record of AI capabilities across business software — what they can do, what they can access, how they are controlled, and how those capabilities change over time.

Capability

What the AI capability actually does and the boundary of the function being recorded.

Data Access

What organisational, user or system data the capability can access or use.

Decision Authority

What judgment or decision the AI itself makes, and where rules, constraints or other systems determine the outcome instead.

Actions

Whether the capability generates information or can cause consequential actions in software, operational or physical systems.

Human Confirmation

Whether an individual action requires human approval before execution, can execute without it, or is not publicly established.

Permission Basis

What permissions, credentials, roles or other authority allow the capability to act, where publicly established.

Administrative Control

How organisations can enable, restrict, configure or disable the capability.

Availability

Which products, plans, regions, deployments or users receive the capability.

Evidence & Verification

What first-party evidence supports the recorded behaviour, including where important details remain not publicly established.

Capability History

How the capability, its controls or its documented behaviour have changed over time.