CrowdStrike launches AI security controls as autonomous agents create new attack vectors

CrowdStrike Falcon Platform Expands to Address AI Security Gaps

CrowdStrike announced a comprehensive expansion of its Falcon platform at the RSAC 2026 Conference, introducing capabilities designed to secure autonomous AI agents as enterprises confront a threat landscape where 48% of cybersecurity professionals rank agentic AI as the top attack vector heading into 20261

1

. The CrowdStrike Falcon platform now establishes the endpoint as the epicenter for AI security, extending AI agent discovery, governance, and AI runtime protection across endpoints, SaaS, browser, and cloud environments2

2

4

4

.

The timing reflects an urgent shift in enterprise security priorities. Only 29% of organizations feel fully ready to deploy AI agents securely, while machine identities now outnumber human employees 82 to 1 in the average enterprise1

1

. IBM's 2026 X-Force Threat Intelligence Index documented a 44% surge in attacks exploiting public-facing applications, accelerated by AI-enabled vulnerability scanning1

1

. CrowdStrike sensors now detect more than 1,800 distinct AI applications running on enterprise devices, representing nearly 160 million unique application instances across its customer base4

4

.

Securing Autonomous AI Agents at the Point of Execution

The company introduced EDR AI Runtime Protection, which delivers runtime visibility of AI behavior at the point of execution. The Falcon sensor captures commands, scripts, file activity, and network connections of all applications running on the endpoint, including agentic applications4

4

. When suspicious behavior is detected, security teams can trace activity to the originating process and act immediately, including isolating affected endpoints to contain threats before they spread2

2

.

CrowdStrike's approach treats the endpoint as the control plane for AI security because AI actions often ultimately execute on a device, making it the optimal place to observe behavior, enforce policy enforcement, and stop malicious activity in real time2

2

. The company also launched shadow AI discovery for Endpoint, which automatically identifies AI applications, AI agents, large language models (LLMs) runtimes, Model Context Protocol servers, and development tools running across devices2

2

4

4

.

Understanding the AI Agent Blast Radius at Machine Speed

Daniel Bernard, CrowdStrike's chief business officer, explained the fundamental difference in the AI agent blast radius compared to compromised human credentials. "Anything we could think about from a blast radius before is unbounded," Bernard said. "The human attacker needs to sleep a couple of hours a day. In the agentic world, there's no such thing as a workday. It's work-always"1

1

.

An AI agent with inherited credentials operates at compute speed across every API, database, and downstream agent it can reach, with no fatigue or shift change1

1

. CrowdStrike's 2026 Global Threat Report documented the fastest observed eCrime breakout at 27 seconds and average breakout times at 29 minutes, but an agentic adversary doesn't have an average—it runs until stopped1

1

. This reality demands oversight architecture that matches detection speed at machine speed.

AI Detection and Response (AIDR) Expands Across Desktop and Cloud

CrowdStrike extended its AI Detection and Response (AIDR) offering beyond browser-based applications to desktop versions, now covering OpenAI ChatGPT, Anthropic's Claude, Google Gemini, DeepSeek, Microsoft Copilot, O365 Copilot, GitHub Copilot, and Cursor3

3

4

4

. The tool delivers real-time prompt inspection and threat detection of prompt injection attacks and data leaks, surfacing access and content policy violations4

4

.

For cloud environments, CrowdStrike introduced Shadow AI Discovery for Cloud, which identifies ungoverned AI services, risky large language models, MCP connections, and sensitive data exposure across infrastructure and application layers2

2

. AIDR for Cloud and Kubernetes brings runtime inspection and enforcement to containerized AI workloads, surfacing malicious activity in Kubernetes environments2

2

. AI Data Flow Discovery for Cloud offers real-time visibility into how sensitive data moves into and through AI services2

2

.

NVIDIA Partnership Delivers Secure-by-Design AI Blueprint

CrowdStrike and NVIDIA unveiled a Secure-by-Design AI Blueprint that integrates protection from the Falcon platform directly into NVIDIA OpenShell, an open-source runtime that enforces policy-based guardrails for autonomous agents5

5

. The NVIDIA partnership marks the first time on a major AI platform release that security shipped at launch rather than being bolted on months later1

1

.

NVIDIA CEO Jensen Huang emphasized the stakes during his GTC keynote: "Agentic systems in the corporate network can access sensitive information, execute code, and communicate externally. Obviously, this can't possibly be allowed"1

1

. The architecture enables organizations to operationalize autonomous systems with governance, visibility, and control from development through runtime, whether agents run locally on NVIDIA DGX Spark or in cloud environments based on the NVIDIA AI-Q Blueprint5

5

.

CrowdStrike and NVIDIA are building what they call intent-aware controls that monitor an agent's planning loop for behavioral drift, creating a different security posture from simple access controls1

1

. The collaboration includes Falcon Endpoint protection on DGX Spark and DGX Station hosts, Falcon Cloud Security across AI-Q Blueprint deployments, and Falcon Identity for identity governance and agent privilege boundaries1

1

.

Microsoft Integration Expands Market Reach for Next-Gen SIEM

CrowdStrike announced support for Microsoft Defender for Endpoint within Falcon Next-Gen SIEM, marking a significant market expansion3

3

. The platform can now ingest and correlate telemetry data from Microsoft Defender for Endpoint, addressing environments where organizations run multiple endpoint security tools3

3

. Bernard called it "another watershed moment for CrowdStrike in the work we're doing with Microsoft"3

3

.

The company also extended AIDR to Microsoft Copilot Studio agents, monitoring prompts, data interactions, and runtime behavior to detect injection attacks, policy violations, and data leaks2

2

. Shadow SaaS and AI Agent Discovery provides visibility into agent activity, permissions, and data access across platforms including Microsoft Copilot, Salesforce Agentforce, ChatGPT Enterprise, OpenAI Enterprise GPT, and Nexos.ai4

4

.

What Security Leaders Need to Watch

The announcements signal a fundamental shift in how enterprises must approach securing autonomous AI agents. Legacy network controls were not designed to govern systems that execute commands, access sensitive data, and trigger workflows autonomously at machine speed with behavior indistinguishable from legitimate user activity4

4

. Michael Sentonas, president of CrowdStrike, stated: "Security built for static applications can't keep up with autonomous systems. Organizations need real-time visibility and control over AI behavior wherever it runs"4

4

.

While CrowdStrike's platform addresses multiple governance layers—from endpoint security for AI to cloud runtime protection and identity-based governance—no single vendor covers all five governance layers identified in the OWASP Agentic Top 10 framework1

1

. Organizations deploying autonomous agents must evaluate whether their security stack can answer critical governance questions across agent decisions, cloud runtime, supply chain provenance, prompt-layer inspection, and pre-production validation. Three or more unanswered vendor questions indicate ungoverned agents in production, creating risk that compounds at the speed of AI execution1

1

.