CrowdStrike brings AI prompt security into focus with Falcon AIDR launch

CrowdStrike has announced the general availability of Falcon AI Detection and Response (AIDR), a new addition to its Falcon platform aimed at securing what the company calls the fastest-growing attack surface in the AI era: the prompt and agent interaction layer. The announcement was made on 15 December 2025, from Austin, Texas.

As enterprises roll out generative AI tools for developers and employees, security teams are grappling with new risks that do not fit neatly into traditional models. Falcon AIDR is CrowdStrike’s answer to this shift, extending security coverage beyond data, models, and infrastructure into the interactions where AI systems interpret prompts, generate responses, and take actions.

Why the AI interaction layer matters

According to CrowdStrike, attackers are increasingly exploiting AI systems through prompt injection and similar techniques. These attacks involve inserting hidden or manipulative instructions into prompts to influence AI behavior, extract sensitive information, or trigger unsafe actions.

“Prompt injection is a frontier security problem,” said Michael Sentonas, president of CrowdStrike, in a statement. He added that adversaries are weaponising the same AI tools that organisations rely on to improve productivity, making real-time protection at the interaction layer a priority.

CrowdStrike frames prompts as the new equivalent of malware, arguing that the interaction layer has become a primary target because it is where AI systems reason and decide what to do next.

Coverage from development to daily use

Falcon AIDR is designed to secure both AI development workflows and everyday workforce usage. For developers, it aims to provide safeguards while building AI-powered applications and agents. For employees, it focuses on monitoring and controlling how AI tools are used across the organisation.

Key capabilities highlighted by CrowdStrike include:

• Visibility into AI usage: Runtime logs show how employees use AI tools and how agents behave, supporting compliance and investigations.

• Blocking prompt injection attacks: Real-time protection against prompt injection, jailbreaks, and unsafe content, informed by research into adversarial prompt datasets and more than 180 known attack techniques.

• Control over risky interactions: Policy-based controls can block unsafe prompts and contain malicious agent actions as they occur.

• Protection of sensitive data: Automated detection prevents credentials, regulated data, and other sensitive information from being sent to models, agents, or external AI systems.

• Support for secure innovation: Built-in safeguards aim to help teams move AI projects into production faster while reducing exposure to new risks.

A unified approach to AI security

CrowdStrike positions Falcon AIDR as part of a broader, unified security model. By integrating AI interaction security into the existing Falcon platform, the company says organisations can manage AI risks alongside endpoint, identity, cloud, and data security from a single platform.

This approach reflects a wider trend in cybersecurity, where vendors are trying to reduce tool sprawl while adapting to the rapid adoption of AI across enterprises.

What comes next

Alongside the product launch, CrowdStrike is promoting additional resources, including blogs, documentation, and a virtual AI Summit scheduled for January 2026 across regions. The focus of these sessions is expected to be secure AI adoption and development.

As generative AI becomes embedded in daily work, the success of platforms like Falcon AIDR may depend on how well they balance control with usability. For now, CrowdStrike is betting that securing prompts and agents will become as essential as protecting endpoints once was.