AI Security
AI Security
AI Security helps you answer a question most organizations can't: which AI tools are actually in use here, and are they safe? It builds an inventory of the AI tools it discovers, highlights shadow AI (tools in use without approval), surfaces the risks they introduce, and lets you set governance policies for how AI may be used.
What it's for
- Discover AI tools in use across your environment and keep them in one inventory.
- Spot shadow AI and risks — unapproved tools, possible data leaks, and policy violations.
- Govern usage — approve or block tools and define policies such as allowlists and usage limits.
The screens in it
AI Security is organized into four tabs:
- Dashboard — the high-level picture: how many AI tools were discovered, how many are shadow AI, how many are approved, and how many risks are open.
- Inventory — every discovered AI tool, with the ability to approve or block each one.
- Risks — shadow AI, data-leak, policy-violation, and unauthorized-model findings.
- Governance — the policies and the list of approved tools.
Every control explained
The AI Security dashboard
| Control | What it shows |
|---|---|
| Discovered AI Tools | The total number of AI tools found. |
| Shadow AI Detected | Tools in use that haven't been approved. |
| Approved Tools | Tools you've explicitly approved. |
| Open Risks | AI-related risks still needing attention. |
| AI Tool Categories | A breakdown by type — LLM services, code assistants, ML frameworks, AI services, and chatbots. |
| Recent Risks | The latest risks detected, with severity. |
| Detection Methods | How tools are discovered — analysis of software dependencies, monitored domain lookups, network traffic patterns, and manually reported tools. |
The Inventory tab
| Control | What it does |
|---|---|
| Search | Filters the list by tool name or vendor. |
| Type filter | Narrows to LLM, Code Assistant, ML Framework, AI Service, or Chatbot. |
| Status filter | Narrows to Approved or Unapproved tools. |
| Approve | Marks an unapproved tool as approved. |
| Block | Blocks a previously approved tool. |
Each row shows the tool's vendor, type, number of users, a risk level, its data-exposure level, whether it's approved, how it was detected, and when it was last seen.
The Risks tab
| Control | What it does |
|---|---|
| Severity tiles | Critical / High / Medium / Low counts. Click a tile to filter to that severity; click again to clear it. |
| Category filter | Narrows to Shadow AI, Data Leak, Policy Violation, or Unauthorized Model. |
Each risk card shows a title, its category, a description, the tool it relates to, when it was detected, and its status — Open, Mitigated, or Accepted.
The Governance tab
| Control | What it does |
|---|---|
| New Policy | Opens a form to create a governance policy. |
| Policy cards | Each shows the policy name, its type, how many rules it has, and whether it's Active, Draft, or Disabled. |
| Approved AI Tools | The list of tools you've approved from the Inventory tab. |
The New Policy form asks for a Policy Name, a Policy Type (Allowlist, Blocklist, Usage Limit, Data Classification, or Approval Workflow), and an Enforcement Mode — Advisory (warn only) or Blocking (deny the action).
Common workflows
Workflow: review and approve a discovered tool
-
Open the Inventory tab. Use the Status filter to show Unapproved tools.
Step 1 — filtering the inventory to unapproved tools -
Review each tool's risk and data-exposure levels and how many people use it.
Step 2 — a tool's risk and data-exposure detail -
Click Approve to allow it, or Block to stop its use. Approved tools appear on the Governance tab.
Step 3 — approving a tool from the inventory
Workflow: create a governance policy
- Open the Governance tab and click New Policy.
- Enter a Policy Name, choose a Policy Type (for example Allowlist), and pick an Enforcement Mode — Advisory to warn, or Blocking to deny.
- Click Create Policy. It appears as a card on the Governance tab.