UK DUAA right to complain enforces in -- days · 19 June 2026 EU AI Act high-risk enforces in -- days · 2 August 2026

Runtime governance

Controls apply before AI-enabled actions proceed.

OBEXGATE turns policies, authority structures, sovereignty limits, and regulatory obligations into runtime decisions.

See how runtime governance differs from model and agent governance →

Configurable enforcement modes.

Allow

Proceed with evidence

The action is permitted and recorded with owner, policy, and evidence context.

Warn

Proceed with visible caution

The action can continue, but the risk signal and warning are captured.

Hold

Pause for review

The action requires reviewer input, SLA handling, or escalation before execution.

Block

Prevent unauthorised action

The action violates a configured control and cannot proceed.

Stop

Terminate unsafe execution

The system stops an action that breaches a hard floor or unacceptable-risk condition.

Override

Controlled exception

Overrides require named authority, reason, limit, and evidence capture.

Admission policies define operating rules.

Governance teams can define decision types, reviewers, override limits, SLAs, escalation paths, and evidence requirements without waiting for vendor bottlenecks.