Autodesk has become the launch customer for Permiso Security’s new AI agent runtime security capabilities. This move aims to provide continuous monitoring and control of autonomous AI agents working across its global workforce, cloud environments, and product ecosystem.
- Permiso's security platform extends to runtime monitoring of AI agents.
- Autodesk leverages this to secure AI-driven activities in cloud and workforce.
- New solution helps detect anomalous behaviors and enforce least privilege.
What happened
Permiso Security has launched new AI agent runtime security capabilities designed to give enterprises continuous visibility into AI agent activity across cloud and on-premises environments. Autodesk is the launch customer for this solution, deploying it across its products, global workforce, and cloud infrastructure to secure AI agents operating with autonomy.
The platform enables security teams to discover all agents operating in their environment, whether managed or shadow, and track their runs, events, tool calls, and data access. It covers various types of AI-related identities including agents, sub-agents, and Model Context Protocol servers, with integrated monitoring of the infrastructure supporting these agents.
Why it matters
Autonomous AI agents are increasingly embedded within enterprise systems, making decisions and interacting with data and tools at machine speed—often without direct human oversight. Traditional identity and security tools primarily focus on agent posture (authentication and permissions), which is static and insufficient for capturing real-time agent behaviors.
Permiso’s approach emphasizes runtime visibility and control, linking every agent action back to a specific human or AI identity. This allows for the detection of anomalous tool usage, over-privileged access, and potential policy violations—critical issues as enterprises accelerate AI adoption. Autodesk’s partnership highlights a growing need for comprehensive security solutions tailored to dynamic AI operations.
What to watch next
Security teams and enterprises deploying autonomous AI agents will likely seek scalable runtime security solutions that provide granular insights into agent activities and enforce stringent behavioral controls. The effectiveness of Permiso’s API-based, agentless architecture without infrastructure changes will be important for market adoption.
Additionally, monitoring how evolving AI security risks—such as LLMjacking attacks and malicious agent skills—are addressed by platforms like Permiso will be critical. Autodesk’s implementation serves as a key case study for how large organizations can secure AI identities as agentic workloads become commonplace.