Exploring security in AI Control Tower

  • Release version: Zurich
  • Updated May 2, 2026
  • 1 minute to read

    • The security area of AI Control Tower offers a dashboard-based view of your AI asset security metrics. In addition to tracking metrics, the security area contains the agent map, a tool that gives a node-graph visualization of the relationships between your AI agents, agentic workflows, and tools.

    Managing AI asset security


    AI Control Tower Security area showing Overview tab metrics.

    Use the security area of AI Control Tower to monitor and improve your security posture across your AI asset ecosystem. Monitor overview, design-time, runtime, and post-runtime metrics to identify problem areas and troubleshoot LLM guardrail policy violations. For details about every metric, see Managing AI asset security reference.

    Security users

    Table 1. Users
    User Description
    AI steward Monitors overall security posture of the AI asset ecosystem. Uses the security dashboard to maintain accountability for AI-related security threats.

    Security benefits

    Table 2. Security benefits
    Benefit Feature Users
    Review and troubleshoot access issues, critical AI asset security events, overall security posture, and other indicators. Overview metrics AI steward
    Evaluate LLM model-level security exposure. Design-time metrics AI steward
    Stop the execution of agents that potentially violate LLM guardrails in real time. Runtime metrics AI steward
    Review AI asset security events to audit AI asset security threats and mitigate risk. Post-runtime metrics and AI agent containment using kill switch protocol AI steward

    What to explore next

    To learn more about configuring and using AI Control Tower, see: