Policies & Guardrails
The policy engine enforces regulatory compliance rules across every platform action, including retention, access, workflow templates, AI usage, and export controls. Policies are treated as operational configuration, so changes can be reviewed and simulated before rollout. This page combines active controls with a simulation panel that returns allow or deny decisions with explicit reasons. Use simulation output to validate persona-based behavior before enabling new workflows.
Active Policies
| ID | Policy Name | Description | Scope | Enforced | Last Updated |
|---|---|---|---|---|---|
| POL-001 | AI Output Review Policy | All AI agent outputs with confidence < 75% require mandatory human review before use. | All AI Agents | Enforced | 2025-01-15 |
| POL-002 | Data Residency Policy | Clinical and patient data must remain within the designated regional data residency zone. | All Modules | Enforced | 2024-11-01 |
| POL-003 | eCTD Publishing Approval Gate | Publishing jobs require an approved snapshot from at least one qualified reviewer before dispatch. | Publishing, Dispatch | Enforced | 2024-09-20 |
| POL-004 | Labeling Change Dual Sign-Off | SmPC/PIL changes affecting sections 4.1, 4.3, or 4.8 require dual sign-off from Labeling and Reg Affairs. | Labeling | Enforced | 2024-08-12 |
| POL-005 | Commitment Escalation Policy | Commitments overdue by more than 14 days are automatically escalated to the Regulatory Affairs Director. | Commitments | Enforced | 2025-02-28 |
| POL-006 | Audit Trail Immutability | Audit events cannot be modified or deleted after creation. Read-only access only. | Audit Log | Enforced | 2023-01-01 |
Policy Simulation
Select a persona and action to simulate allow/deny behavior with policy reasoning.
Persona
Action
ALLOW - Action is permitted by current policy template and will be fully audited.