Administration
This section is for administrators — security, compliance, and IT — who configure how ThreatLens governs AI for the organization. It's the largest part of the documentation, because it's where the product's value is set.
New to the model behind these settings? Start with Core concepts.
Guides in this section
- Configure the policy matrix — set, per data class, the minimum trust tier, the fallback action, and the internet policy.
- Connect models and trust tiers — add AI providers and assign each a trust tier, including your own Azure OpenAI / AWS Bedrock (BYOK).
- Data-leak protection and rules — review what's detected and add custom rules.
- Monitor and enforce — baseline usage, then make policy binding.
- Connect Microsoft 365 — enable governed grounding from OneDrive and SharePoint.
- The secrets vault — store provider and connector credentials securely.
- Single sign-on — SAML, Microsoft Entra ID, and Google Workspace, with group-to-role mappings.
- Roles and access (RBAC) — control who can see and change what.
- The audit and governance log — read, filter, and export the complete decision record.
A recommended rollout
- Connect identity (SSO) and define roles.
- Connect your enterprise-managed model(s) and assign trust tiers.
- Review the policy matrix and adjust it to your risk posture.
- Run in monitor mode to baseline real usage.
- Switch to enforce mode to make policy binding.
- Connect Microsoft 365 for governed grounding.
Everything here is audited
Every change you make in this section is recorded in the audit log, with the old and new values. The people who govern AI are themselves governed.