The news: Cyata launched a platform that detects, authenticates, and governs “agentic identities” as adoption of autonomous AI agents is exploding—96% of IT leaders will increase agent use in 2025, Cloudera reports. 

Digital agents integrating into the workforce pose new risks—ones traditional identity and access management (IAM) tools are not equipped to handle, per VentureBeat.

“Enterprises need new guardrails to handle the velocity and autonomy of these systems,” said Shahar Tal, Cyata co-founder and CEO. 

What Cyata offers:

• Autonomous detection: Identifies AI agents across hybrid environments without manual tagging.
• Real-time observability: Monitors agent actions live, not postmortem.
• Least privilege enforcement: Limits agent access to only what’s necessary, reducing risk surface.
• Audit-ready logs: Tracks agent behavior for compliance and forensic investigations.

The tipping point: AI agents are rapidly gaining ground in business, boosting efficiency across customer experience, data analysis, and project management. Consumers are warming up to them, too—34% are comfortable letting agents handle research, and 22% trust them with task automation, according to Auth0.

Solving a new problem: While security is often an afterthought, especially in AI, solutions like Cyata can anticipate AI agent integration pain points. 

• It assigns each AI agent an agentic identity that can be easily managed through automated agent discovery, real-time observability, and granular access control across enterprise systems.
• “We identify agents at login, track their actions, and enforce … the minimum access needed to perform its job,” Tal said. 

Possible caveats: Integrating new identity layers may slow down teams. Enterprises need cultural buy-in to trust software guarding software, and large-scale deployments with thousands of agents remain untested in public.

Our take: Managing mixed human and agentic workers won’t be optional for long—it will become a baseline requirement as AI agents move from edge cases to everyday tools. Companies that delay could risk operational blind spots, compliance gaps, and uncontrolled AI autonomy.