§ 08 · Identity analytics — finding hidden risks
Powered by Radiant Logic

The system finds risks reviews miss.

Peer-group clustering surfaces specific, actionable findings — not vague "something looks wrong" alerts that auditors must investigate by hand.

Peer-group clustering

21 natural groups · 64 employees

Employees are clustered by similar access patterns. Anyone who doesn't fit is flagged automatically.

Example: an HR specialist who also holds FX trading access is immediately visible.

Outcome

25 of 64 flagged

Each finding includes the specific risk factor and a recommended action — not a single opaque score.

What contributes to a risk score

  1. Permissions no peer holdsOutlier access compared to peer group.
  2. Cross-department accessSystems outside the employee's home department.
  3. Above-average permission countMore entitlements than peers in the same role.
  4. High-risk concentrationDisproportionate share of high-risk permissions.
  5. Inactive deviceLast seen 30+ days ago — possible stale account.
  6. Orphaned grantAccess granted by a manager no longer in role.
Previous07 · PEP — The enforcement layer