Okta Acquires Permiso to Tackle the Enterprise AI Agent Security Crisis
Okta is acquiring AI identity security startup Permiso to bolster its ability to monitor non-human identities, as enterprises increasingly deploy autonomous software agents across their cloud environments.
Okta has agreed to acquire Permiso Security, an identity security startup, in a move to address the growing challenge of securing machine identities and autonomous AI agents. According to TechCrunch, the acquisition is valued at just under $200 million and is structured as an almost all-cash deal. The transaction is expected to close in the third quarter of Okta’s fiscal 2027. This integration aims to expand Okta’s existing identity management framework, moving beyond simple login authentication to provide continuous monitoring of what applications and AI agents do once they have accessed a network.
Permiso, which emerged from stealth in 2022, was founded by former FireEye executives Paul Nguyen and Jason Martin. The startup specializes in detecting attacks that leverage compromised identities to move laterally through cloud infrastructure. Its platform analyzes the behavior of API keys, roles, and automated agents to identify suspicious activity. More recently, the company introduced a tool called SandyClaw, which evaluates the skills of AI agents within a sandboxed environment to detect potential malicious behavior before those agents are deployed into production systems.
The acquisition highlights a significant shift in enterprise security architecture. As businesses embed AI deeper into their operations, the security perimeter has expanded beyond human employees to include a vast, multiplying fleet of non-human identities. Ely Kahn, Okta’s chief product officer, stated that Permiso will extend the company’s identity security fabric with proven threat detection and response capabilities, while bringing in a specialized research and security team to advance Okta’s overall prevention efforts.
While Okta focuses on this integration, the broader market for AI-related security continues to evolve. Other startups are attracting capital to address similar gaps in the enterprise landscape. For instance, London-based Inforcer recently raised $50 million to develop tools for managed service providers, including features to detect unauthorized use of AI tools on company devices. These market moves underscore a collective industry recognition that traditional defense models, designed primarily for human users, are insufficient for the current era of automated enterprise workflows.
Permiso previously raised approximately $29 million in funding, including an $18.5 million Series A round in April 2024 led by Altimeter Capital. The company’s focus on runtime behavior analysis addresses a specific vulnerability: automated entities often possess over-privileged access to sensitive databases, making them prime targets for attackers. By mapping session activity and blocking anomalous behavior in real time, Permiso provides a layer of defense that Okta intends to leverage as it seeks to secure the next generation of automated enterprise software.