Admin dashboard — one pane of glass for AI governance
Built for the security, compliance, and risk teams responsible for what happens when 10,000 employees type into a text box. Sign in, see everything, prove it to auditors.
Live risk score
A composite score blending detection volume, severity, and framework exposure across the organisation. Trends over time surface in a single glance.
Shadow-AI discovery
See which AI tools your employees actually use, grouped by department, platform, and region — including usage your MDM missed.
High-severity queue
A triage queue for critical detections. Each event links to its category, user identifier, and framework tag — no prompt content, ever.
Policy controls
Turn rule families on or off per region, per tier, per department. Changes are version-controlled and logged in the audit trail.
SSO and RBAC
SAML / OIDC single sign-on, fine-grained roles for Admin, Security, Compliance, and Read-only views. Service accounts for your SIEM pulls.
Evidence exports
Every view is exportable as CSV or signed PDF for your auditors, with stable identifiers so evidence ties back to a specific detection event.
Role-aware by design
A CISO and a DPO don't want the same screen. Admin Dashboard surfaces different defaults per role: security gets the high-risk queue and SIEM integrations on open; compliance gets the framework heatmap and audit exports; finance gets seat usage and trial status. Everyone sees only what their role permits.
Built on a zero-knowledge substrate
Because detection runs in the browser, the Dashboard never has access to prompt content. What it does have is rich, structured metadata — category, severity, framework, platform, time, user — which is everything you need for risk reporting, and nothing more. This is the difference between a compliance tool and a surveillance tool.
Integrations
- SIEM: Splunk, Datadog, Elastic — push detection events via outbound webhook or pull via REST.
- Identity: Okta, Azure AD, Google Workspace via SAML and OIDC.
- Ticketing: Jira and ServiceNow for high-severity events that require human follow-up.
- MDM: Deployment templates for Jamf, Intune, and Kandji to roll out the extension at scale.