AI Security
Findings Dashboard
When a Glean AI security policy is violated, the incident is logged and displayed on the Findings dashboard. This page provides a detailed list of every flagged or blocked incident, helping administrators review potential threats to their AI agents.
To access the dashboard, in the Glean Admin Console, navigate to Glean Protect > AI security and click the Findings tab.
Additionally, you can click on the three-dot menu next to each finding to access options such as copying the unique Run ID or viewing the specific prompt and response in a Show snippet view.
Findings Details
Each entry on the dashboard provides a snapshot of a violation with the following information:| Field | Description |
|---|---|
| Agent | The specific agent (e.g., Glean Chat, Summarize) where the violation was found. |
| Policy Violated | The name of the security policy that was triggered (e.g., “Prompt injection attempt in user prompt”). |
| Status | The action taken by the system. A Blocked status means the request was stopped. A For review status means the incident was flagged for administrator review but the agent run was allowed to proceed. |
| Detected | The timestamp indicating when the violation was detected. |
| User | The user who initiated the agent run. |