This design assumes a centrally governed security operating model where all AI interactions are identity-authenticated and policy-controlled with no implicit trust boundaries. Ownership is consolidated within a security platform team enforcing uniform access and execution discipline, while consumption is restricted to governed pathways. Execution is anchored in Azure through Microsoft Entra ID and Azure Policy, ensuring identity-led access control and policy-driven behaviour across all AI workloads. The model prioritises continuous verification, controlled autonomy, and auditable operations over permissive access.