Dashboard
(past 30 days)→
Open findings
19
Critical 3High 8Medium 6Low 2
→Shadow AI app count by severity
114
Critical 9High 14Medium 56Low 35
→Sanctioned vs. Unsanctioned AI Use
7%
Sanctioned
Sanctioned 8Unsanctioned 106
→Coarse vs. Fine-Grained Rule Hits
61%
Coarse
Coarse 44Fine-grained 28
Findings by use casei
LowMediumHighCritical
Top rules triggering mediation
| Rule | Use case | Hits |
|---|---|---|
| 1No agent may call execute_shell | Tool-Call Mediation | 21 |
| 2Reject messages relayed through an untrusted/unknown hop | Agent-to-Agent | 21 |
| 3query_database restricted to allow-listed tables | Tool-Call Mediation | 18 |
| 4send_email flagged for recipients outside company.internal | Tool-Call Mediation | 10 |
| 5Per-visitor scan ceiling (5/hour) | Content Broker | 1 |
Risk score over time
Mediation activity over time
BlockRedactFlag
Top Risky Shadow AI Users
| User | Dept | Findings |
|---|---|---|
| 1j.alvarez | Marketing | 9 |
| 2t.okafor | Engineering | 7 |
| 3m.chen | Sales | 6 |
| 4r.patel | Finance | 5 |
| 5s.kowalski | Engineering | 4 |