01 / capability
Agent authority and role policies
Defined as an explicit contract so operators can review it before it runs.
Authority & assurance · Prototype
Authority before action.
An enterprise Agent Control Plane that authorizes, governs, supervises and audits actions performed by AI agents and automated systems.
01
What it is
EnterpriseACP is built for CISOs, CIOs, enterprise AI platform teams, governance leaders.
The problem
EnterpriseACP replaces scattered agent permission configurations, manual approval systems and application-specific security controls.
The approach
An enterprise Agent Control Plane that authorizes, governs, supervises and audits actions performed by AI agents and automated systems.
The boundary
Work stays inside an explicit scope: named sources, defined permissions, bounded execution, retained evidence.
02
Capabilities
Five contracts the product is specified around. Each one is a boundary as much as a feature.
02 / capability
Pre-execution authorization
Defined as an explicit contract so operators can review it before it runs.
03 / capability
Approval and delegation workflows
Available through the product surface and the machine interfaces, with permission attached.
04 / capability
Budgets, limits and revocation
Exposed as a scoped capability with its own evidence trail, not a hidden setting.
05 / capability
Execution evidence and audit
Available through the product surface and the machine interfaces, with permission attached.
03
How it works
The same three steps apply whether a person or an agent is driving.
01
Define the scope
Start from the boundary EnterpriseACP is allowed to act inside: the workspace, the sources, the budget, the authority.
02
Connect the surfaces
Expose the same capability through the interface a human uses and the interface an agent calls, with one permission model.
03
Act with evidence
Every action leaves a record that can be inspected, replayed and challenged before the next one runs.
Read the specification before you believe the category.
EnterpriseACP is documented as a prototype: requirements, interfaces and evidence expectations. It is not presented as a live service.