CrowdStrike Enhances AI Agent Security Across SaaS with Falcon Shield Integration for ChatGPT Enterprise

CrowdStrike has expanded its AI agent security capabilities with the introduction of new functionality in Falcon Shield, designed to provide visibility, governance, and protection for AI agents across the SaaS stack. The latest update integrates with OpenAI’s ChatGPT Enterprise Compliance API, enabling organizations to monitor and manage GPT-based agents built using ChatGPT Enterprise and OpenAI Codex. This enhancement extends support to more than 175 SaaS applications, including platforms like Microsoft 365, Salesforce, and Snowflake. As AI agents become increasingly embedded in daily workflows, they introduce new security risks. These autonomous, non-human identities often have persistent access rights and can be exploited if the human account that created them is compromised—potentially leading to data exfiltration, system manipulation, or lateral movement across critical systems. Falcon Shield now discovers AI agents across these platforms, maps them to their human creators, detects risky behaviors, and enables automated threat containment. By linking each agent to its originator, security teams gain accountability and context, which is essential for enforcing policies and managing access. The solution also leverages Falcon Identity Protection to secure the underlying human identities, ensuring that the entire identity lifecycle—from creation to access—is monitored and protected. Key capabilities of Falcon Shield include: Discovering AI agents across SaaS environments, including custom GPTs and Codex-powered tools. Mapping each agent to its human creator, enabling traceability and governance. Identifying high-risk behaviors such as overprivileged access, unusual actions, or access to sensitive data. Automatically containing threats using Falcon Fusion, CrowdStrike’s no-code SOAR engine, which can block access, disable compromised agents, or trigger response workflows. Delivering unified protection across human and non-human identities through integration with Falcon Identity Protection and Falcon Cloud Security. Elia Zaitsev, CrowdStrike’s Chief Technology Officer, emphasized the growing threat posed by AI agents. “AI agents are emerging as superhuman identities—capable of executing complex tasks at machine speed. As they multiply across SaaS environments, they’re reshaping the attack surface. Their security depends entirely on the integrity of the human identities behind them. Falcon Shield and Falcon Identity Protection help secure this new layer of risk.” The announcement was made at Black Hat USA 2025 in Las Vegas, highlighting CrowdStrike’s role as a leader in securing the AI-driven enterprise. The company continues to evolve its Falcon platform to address the expanding scope of identity-based threats in cloud and SaaS environments. For more information on how CrowdStrike secures AI agents across the SaaS stack, visit the company’s blog or stop by booth #2733 at Black Hat.