CrowdStrike launches Falcon Guardian for AI agents
Wed, 2nd Sep 2026 (Today)
CrowdStrike has launched Falcon Guardian, an AI detection and response product for AI agents, and expanded Falcon integrations across Google Cloud's enterprise AI tools.
The move adds a security offering focused on monitoring and controlling AI agents at the endpoint, while extending CrowdStrike's software into Google Agent Gateway, Gemini Enterprise and Google Cloud's Agent Registry.
Falcon Guardian is designed to give security teams visibility into AI agents running across corporate environments, including endpoints, cloud services, software-as-a-service applications and browsers. The product covers data, models, prompts, agents, identities, infrastructure and interactions, reflecting the broader set of assets companies must manage as AI tools move from pilot projects into day-to-day operations.
The launch centres on the endpoint as the primary enforcement point. CrowdStrike argued that AI agents increasingly operate with system-level access and can carry out actions that resemble legitimate user behaviour, making it harder for organisations to rely on governance rules alone once an agent is active.
New Falcon Guardian functions include AI Agent Discovery and Inventory, which can identify known and shadow AI agents on Windows and macOS devices, and Agent Access Controls, intended to determine which agents may run on managed endpoints. The product also includes runtime detection and response capabilities, as well as an AI Gateway designed to act as a central point for AI traffic across supported models and services.
Falcon Guardian will also connect with Falcon Next-Gen SIEM so AI agent data can be correlated with identity, cloud and software-as-a-service information. Managed services linked to the product will include Falcon Complete for Guardian and Falcon Adversary OverWatch for Guardian.
George Kurtz set out the company's view of how AI changes the security problem.
“CrowdStrike pioneered EDR by making the endpoint the control point for stopping attacks. AI demands the same approach,” said George Kurtz, Chief Executive Officer and founder, CrowdStrike.
“AI hasn't changed the attack, it has changed its speed. Governance alone can't stop an agent already in motion. Falcon Guardian turns policy into protection, stopping threats where AI agents execute and before they can cause harm,” said Kurtz.
Google Cloud tie-up
Alongside the product launch, CrowdStrike is broadening Falcon's role across Google Cloud's enterprise AI environment. The expansion brings Falcon Guardian into Google Agent Gateway, where it will provide runtime protection for enterprise AI applications built on Google Cloud.
The integration is aimed at risks such as prompt injection, sensitive data leakage and malicious AI activity. It is intended to give customers continuous visibility across AI agents and applications as they build and run AI services on Google Cloud.
CrowdStrike is also extending Falcon MCP, Charlotte AI and Falcon Shield into Gemini Enterprise. According to the company, this will bring threat intelligence, detections and other security context into workflows using Google's Gemini models, while supporting governance through Google Cloud's Agent Registry.
CrowdStrike added that it is building the Falcon platform on regional Google Cloud infrastructure, a step intended to help customers standardise on CrowdStrike while aligning with cloud deployment preferences and operational requirements.
Daniel Bernard said the partnership reflects rising demand for security tools that sit alongside enterprise AI roll-outs rather than around them.
“Enterprise AI needs an enterprise cybersecurity platform,” said Daniel Bernard, Chief Business Officer, CrowdStrike.
“Organisations shouldn't have to choose between accelerating AI adoption and reducing risk. Through our collaboration with Google Cloud, we're extending the Falcon platform across the enterprise AI stack, helping customers run AI securely at scale in the location of their choice,” said Bernard.
Broader shift
The announcements come as cybersecurity suppliers seek to define how security teams should monitor autonomous and semi-autonomous AI systems. Unlike traditional software, AI agents can initiate actions, call tools, move data between systems and interact with business workflows with limited human intervention, raising questions about runtime oversight, access control and forensic visibility.
CrowdStrike's approach places that oversight at the device level while linking it to the cloud and software-based environments where AI agents may also operate. The company is effectively arguing that conventional governance and posture management are not enough on their own once AI tools begin acting in live production settings.
Google Cloud also framed the relationship as part of a wider push to embed security more deeply into enterprise AI deployments.
“Enterprise AI is rapidly becoming a new operating layer for the enterprise, making security foundational to every AI application and agent,” said Brian Goldstein, Vice President, Strategic AI and ISV, Google Cloud.
“Integrating CrowdStrike's AI security capabilities with Gemini Enterprise further strengthens our native protections, helping customers scale enterprise AI on Google Cloud with comprehensive, defense-in-depth security,” said Goldstein.