The pipeline

Graph in, remediation out - inside your tenant

Four stages, all running in your own Azure subscription. Nothing is crawled from endpoints and no copy of your data is shipped to a vendor cloud. Admins run each stage on a schedule or on demand, and can pause a long sync at a safe checkpoint.

Connect

One Entra app registration with read-only Graph permissions, deployed in your subscription.

  1. 1Register in Entra ID
  2. 2Grant read-only scopes
  3. 3Deploy to your Azure

Synchronize data

The worker syncs Entra ID, SharePoint, OneDrive, sharing, Copilot Studio telemetry, and Copilot usage - reading only what changed.

  1. 1Graph delta sync
  2. 2Agent telemetry & status
  3. 3Copilot usage reports

Recalculate insights

Deterministic scoring and prompt classification over your synced data, with an optional AI narrative on top.

  1. 1Score sites & agents
  2. 2Classify every prompt
  3. 3Snapshot daily posture

Act & report

The Executive Summary, alerts, and PDF reports surface the queue; admins remediate, and everything is logged.

  1. 1Review alerts
  2. 2Revoke, block, quarantine
  3. 3Email PDF reports

Entra ID single sign-on

Staff sign in with their existing work account. Access is scoped to the Entra ID groups you choose, with separate roles for admins, users, and data stewards, re-checked against Entra as you work.

Runs in your own tenant

Deployed into your Azure subscription. Inventory lives in your PostgreSQL database and never leaves your environment. There is no shared multi-tenant backend.

Read-only by default

Sentinel connects to Microsoft Graph with least-privilege, read-only permissions. Remediation actions - like revoking a link - happen only when an admin starts them.

Full audit trail

Every revoke, bulk revoke, block, and quarantine is logged with actor, target, outcome, and timestamp, and exportable as CSV. Nothing is used to train public models.

Secrets stay in Key Vault

Configuration and signing keys are read from Azure Key Vault through managed identity. An admin's delegated sign-in is sealed with AES-256-GCM before it is stored.

Hardened by default

Sessions end after 12 hours, every form is CSRF-protected, pages ship a strict content security policy, and the worker API is never exposed to the internet.

Location restriction

Optionally limit sign-in to the countries you allow, with an Entra Conditional Access policy that starts in report-only mode and can be rolled back.

Minimal data kept

Microsoft 365 Copilot analytics store prompt counts, never prompt text. Copilot Studio message content is removed after 90 days, and profile photos are never stored.

Deployment & data residency

Your tenant, your data, your control

Princeton Sentinel ships as containers on Azure Container Apps - a Next.js web app and a Python worker - with a dedicated Azure Database for PostgreSQL, Azure Key Vault for configuration, and Application Insights for monitoring, all deployed into your own subscription. Sign-in and role membership run through your Microsoft Entra ID, and AI summaries are optional: reuse an Azure OpenAI deployment you already have, create one, or leave it out entirely. There is no shared multi-tenant backend, upgrades are applied in place, and nothing is ever used to train public models.

Editions

Start with exposure, grow into AI governance

Every deployment is licensed per tenant with a signed license file, so features switch on without a redeploy. Trials include the full Enterprise feature set, and if a license lapses your synced data stays readable while syncs and actions pause. See what the AI governance suite covers.

01

Essentials

Know your exposure: Microsoft Graph sync, the Executive Summary and SharePoint, OneDrive, and Entra ID dashboards, Data Risk scoring, and scheduled email reports.

02

Professional

Fix your exposure: everything in Essentials, plus permission remediation - revoke specific links, bulk revoke, and the Revoke Activity audit log.

03

Enterprise

Govern data and AI: everything in Professional, plus the AI governance suite - custom agents, Agent Risk, AI Safety, agent quarantine, Copilot Utilization, and AI narratives.

Have a security questionnaire?

Bring it to the demo - we'll walk your IT and security teams through identity, data flows, permissions, and hardening in the same session.

Request a demo