Use-case library
The tool is deliberately lightweight so it can be used by AI governance teams, researchers, auditors, product teams, public-sector teams, IoT teams, and agent workflow developers.
Domain: Agentic AI / workplace automation
Proposed action: An AI assistant sends a contract-related message to a customer.
Governed effect: External communication that may create institutional commitment or reputational risk.
A fluent AI-generated message can look harmless but still require mandate, constraints, and reviewable proof.
Domain: Public sector / administrative decision support
Proposed action: A system recommends issuing, withholding, or escalating a citizen-facing decision.
Governed effect: A public administrative effect that may affect rights, access, or obligations.
The tool distinguishes model recommendation from authority to produce a governed effect.
Domain: IoT / cyber-physical systems
Proposed action: A cloud system sends a command to unlock, shut down, restart, or change a device state.
Governed effect: A physical or operational state change.
Live-context integrity and proof are needed before a digital instruction becomes a physical effect.
Domain: Robotics / embodied AI
Proposed action: A robot performs a movement, delivery, inspection, or intervention task.
Governed effect: A real-world action affecting people, assets, or environment.
The execution boundary should check authority, context, safety constraints, accountability, and evidence.
Domain: Enterprise AI / customer operations
Proposed action: A system sends a fraud warning, service suspension notice, or payment instruction.
Governed effect: Customer-facing communication that can trigger behaviour, loss, or dispute.
The action should be reviewable and evidence-aware before it reaches the customer.
Domain: Institutional governance / procurement
Proposed action: An AI-supported workflow recommends approval of a supplier, purchase order, or payment release.
Governed effect: A financial or institutional commitment.
Valid constraints, accountable ownership, and verifiable evidence reduce governance and audit risk.