CrowdStrike to acquire Fiddler for AI agent observability

Securing the agentic enterprise

CrowdStrike has announced its intent to acquire Fiddler, a company specializing in AI observability. This move is designed to help businesses manage the risks associated with AI agents, which are software systems capable of making autonomous decisions and using enterprise tools. Unlike traditional software, these agents can act independently across various systems, creating new challenges for security teams.

The primary concern for many organizations is not just whether an AI system is compromised, but whether it makes a bad decision using valid credentials. For example, an agent might have the permission to process a customer refund but could incorrectly approve one that violates company policy. Because the activity appears legitimate to traditional security controls, companies need a way to see the reasoning behind the action.

By integrating Fiddler into the Falcon platform, CrowdStrike aims to provide a way to trace what each agent decided, how it acted, and which tools it used. This level of visibility is becoming a priority for auditors, regulators, and risk managers who need to understand the logic behind automated actions. The goal is to move beyond just protecting the AI system to proving that the agents are behaving as intended every day.

How AI observability works

AI observability provides a lens into how an agent behaves both before and after it is deployed into a production environment. Fiddler specifically focuses on AI explainability, which shows why an agent made a certain decision and where its behavior might have gone wrong. This allows teams to identify misaligned agents before they reach the public or internal users.

Before deployment, teams can use the technology to compare different prompts, models, and configurations. This evaluation process helps ensure that the agent performs its intended tasks accurately and safely. By testing these variables early, organizations can catch errors in the agent's logic before the software is actually shipped to the workforce.

Once an agent is live, the tool continues to monitor its behavior to detect issues like behavioral drift. Behavioral drift occurs when an agent's actions begin to deviate from its original programming or intended purpose over time. Fiddler reconstructs the decision-making process so that teams can trace a specific failure back to the exact step that caused the error.

Integration with Falcon platform

The acquired technology will be rebranded as CrowdStrike Falcon Observe and is expected to be generally available in Q4 FY27. It will connect the intent and behavior of AI agents with existing security telemetry across endpoints, identity, and cloud environments. This integration allows security teams to see the full context of an agent's actions across the entire enterprise.

Fiddler supports several major frameworks, including LangGraph, Amazon Bedrock, AWS Strands, and Google ADK. It also enables teams to instrument agents through OpenTelemetry, a standard for collecting logs and metrics. This broad compatibility ensures that the tool can work with various first-party and third-party agents already in use by a company.

By combining these capabilities, CrowdStrike is building what it calls an enterprise AI control plane. This structure connects the human operators of AI agents to the agents themselves, granting access based on real-time risk. It allows for a comprehensive view where security teams can understand not just that an action happened, but why it happened and what the result was across the network.

Governance and compliance

A major component of this acquisition is the ability to support auditable AI governance. Because AI agents can invoke tools and take actions across enterprise systems, there must be a record of those actions for compliance purposes. Fiddler provides a granular view of these decisions, which is essential for meeting the requirements of modern regulators and internal board members.

The platform helps organizations identify and fix issues where agents might be taking harmful actions using approved tools. By providing a clear trail of reasoning, the software helps turn

The path to AI safety

CrowdStrike is positioning this acquisition as a milestone in protecting the

Next steps for AI safety

CrowdStrike intends to combine Fiddler's capabilities with its existing Falcon Guardian innovations to create a complete safety suite. While Falcon Guardian focuses on protecting AI agents from external attacks and misuse, Falcon Observe will be used to evaluate and explain whether the agents are behaving as intended. Together, these tools aim to secure the entire lifecycle of AI operations.

The acquisition is currently an announcement of intent, and the final integration will involve several steps to ensure the technology works seamlessly within the Falcon platform. The company notes that the success of this move depends on the ability to integrate Fiddler effectively and meet various closing conditions. For now, the focus remains on providing the visibility needed to build trust in autonomous systems.

As businesses continue to deploy more agents, the demand for these types of tools is expected to grow. Organizations will need to balance the benefits of automation with the necessity of oversight. By providing a way to reconstruct decision paths, CrowdStrike is attempting to give risk managers the tools they need to manage this transition safely.