CrowdStrike CEO George Kurtz said AI agents must be secured, controlled, and governed to protect the evolving digital enterprise [1].
This push comes as AI agents move into production faster than security teams can monitor them. Without oversight, these systems create critical vulnerabilities, including prompt-injection attacks, data leakage, and manipulation by adversaries [4, 5].
Speaking on CNBC’s Closing Bell Overtime, Kurtz said there is a need for visibility and control across the AI interaction layer [1]. To address these gaps, CrowdStrike is adding identity and control capabilities to its portfolio. This includes the introduction of Continuous Identity for AI Agents, which Kurtz said reinforces the Falcon platform as the identity security control plane for the agentic enterprise [3].
The company is also expanding its reach through strategic alliances. CrowdStrike is partnering with firms such as AWS and Salesforce to provide a comprehensive AI security layer [2, 6]. These collaborations aim to standardize how businesses secure the future of AI-powered operations [6].
Kurtz said the demand for these solutions is scaling rapidly. He said thousands of partners [2] are looking to the company for answers on how to secure AI at a global scale [2].
These initiatives were highlighted during recent appearances at the Identiverse 2026 conference in Las Vegas and in interviews this month [1, 3]. The company is positioning itself to act as the primary security layer for the autonomous agent ecosystem, focusing on the intersection of identity and AI governance [2].
“AI agents need to be secured, controlled, and governed.”
The shift toward 'agentic' AI—where AI doesn't just answer questions but takes actions—creates a new attack surface. By focusing on identity and governance, CrowdStrike is attempting to move beyond traditional endpoint protection to manage the permissions and behaviors of autonomous software, effectively treating AI agents as digital identities that require the same scrutiny as human users.
