AI agent runtime control: what security teams need now

TL;DR: AI agents in enterprise environments created risk through inherited permissions, tool invocation, data movement, and runtime behaviour, according to Zenity’s 2025 year-in-review. The governance gap is that traditional IAM and monitoring models were built for stable identities and not for agents that decide and act during execution.

NHIMG editorial — based on content published by Zenity: Zenity 2025 Year in Review, Building AI Security for the Enterprise

By the numbers:

• 72% of organisations have experienced or suspect they have experienced a breach of non-human identities - 46% confirmed, 26% suspected.

Questions worth separating out

Q: How should security teams govern AI agents that inherit enterprise permissions?

A: Treat the agent as a non-human identity with runtime behaviour, not just as a model wrapper.

Q: Why do AI agents complicate traditional IAM and PAM controls?

A: Because traditional IAM and PAM assume access is relatively stable and reviewed over time.

Q: What breaks when agent visibility is not paired with runtime enforcement?

A: Teams can see that an agent exists and still fail to stop unsafe behaviour.

Practitioner guidance

• Inventory every agent identity and tool path Map where agents operate across SaaS, cloud, and endpoint environments, then record which authenticated sessions, APIs, and tools each one can reach.
• Move control points to execution time Require policy checks when the agent attempts an action, especially where data could be sent to an unapproved domain or a tool could change state in a critical system.
• Correlate posture with runtime evidence Join entitlement data, data-flow context, and runtime anomalies into one investigation workflow so analysts can see how an issue unfolded.

What's in the full article

Zenity's full year-in-review covers the operational detail this post intentionally leaves for the source:

• Environment-by-environment coverage for SaaS-managed, home-grown, and device-based agents
• The Issues and Correlation Agent workflow for reconstructing agent incidents across multiple signals
• Inline Prevention behaviour across Copilot Studio, Azure AI Foundry, and endpoint environments
• Tool-layer coverage for MCP servers and the policy patterns used to block unsafe destinations

👉 Read Zenity's year-in-review on building AI security for the enterprise →

AI agent runtime control: what security teams need now?

Explore further

View Full Forum →  |  NHI Foundation Course →