CrowdStrike to Acquire Fiddler to Deliver Falcon Observe for AI Agent Observability

Fiddler will become part of the Falcon platform as Falcon Observe, generally available in Q4 FY27.

I am delighted to announce CrowdStrike’s intent to acquire Fiddler at Fal.Con EU 2026. Fiddler, a leader in AI observability, will help ensure enterprise AI agents act as intended with pre-deployment evaluation and continuous behavioral monitoring as part of the CrowdStrike Falcon® platform.

Every business deploying AI agents asks the same questions: What are our agents doing? Are they staying on task? Are they acting in a way that could put the business at risk?

AI agents don’t have to be compromised to be dangerous. Those that stay within their guardrails can make a bad decision or take harmful action using valid credentials and approved enterprise tools. Businesses must be able to trace agents’ activity and explain anomalies as these concerns become top of mind among auditors, regulators, board members, and risk managers.

Enterprise AI safety requires a comprehensive approach: Evaluate agents before deployment, understand their decisions and actions, and continuously monitor their behavior. AI observability provides a lens into agent behavior before and after they ship. AI explainability shows why agents made their decisions and where their actions may have gone awry.

Fiddler delivers both. It will become part of the CrowdStrike Falcon® platform as CrowdStrike Falcon® Observe, generally available in Q4 FY27. With this acquisition, CrowdStrike will extend our AI security capabilities beyond protecting AI systems into proving businesses can safely use AI agents every day.

Building AI Trust in the Agentic Enterprise

Unlike traditional software, AI agents can make autonomous decisions, invoke tools, and take actions across enterprise systems. If software can act autonomously, security needs a record of those actions: what each agent decided, how it acted, what it changed, and the tools it used. They must know whether the agent made the correct decisions to follow its intended task.

As an example: Consider an AI agent using approved tools and valid credentials to process a customer refund. The agent makes an incorrect decision and approves a refund that violates company policy. Traditional security controls may see legitimate activity, but AI explainability reveals why the agent made that decision and where its behavior went wrong.

Fiddler provides AI observability for the agentic era. It helps organizations identify misaligned agents before they are deployed and continues to monitor agent behavior post-deployment so businesses have full visibility into what they’re doing and why.

Before deployment , teams can compare prompts, models, and configurations to evaluate whether agents perform intended tasks safely and accurately before they reach production.

After deployment , teams can trace agents’ decisions and reasoning paths across the full agent hierarchy to detect issues like behavioral drift and unintended actions before they cause harm. Fiddler reconstructs the agent’s decision-making process so teams can trace each failure to the step that caused it.

Fiddler supports auditable AI governance across organizations’ existing agents and tools. Its granular view of agents’ decision-making covers first-party and third-party agents across the enterprise so teams have the data they need to support compliance and audits. Businesses also gain explainability across the frameworks they run, as Fiddler enables teams to instrument agents through OpenTelemetry and supports LangGraph, Amazon Bedrock, AWS Strands, and Google ADK.

As part of the Falcon platform, Fiddler will connect AI agent intent and behavior with Falcon security telemetry across endpoints, identity, cloud, and runtime environments. This will help security teams understand why an agent took an action and what happened across the enterprise as a result. By connecting agent decisions with enterprise security context, CrowdStrike will provide visibility that goes beyond standalone AI observability.

CrowdStrike Delivers the Enterprise AI Control Plane

This milestone extends CrowdStrike’s leadership in protecting the agentic enterprise. Our acquisition of Fiddler, together with CrowdStrike Falcon® Guardian innovations announced today, will bring comprehensive AI security and safety solutions to the Falcon platform.

CrowdStrike can already establish AI agents as trusted identities, connect those identities to the humans operating them, grant access to business systems based on real-time risk, and secure AI agents at runtime. Fiddler will extend the Falcon platform to the earliest stages of the agent lifecycle, before deployment, to monitor their behavior before and after production, and explain the decisions they make.

Falcon Guardian protects AI agents from attacks and misuse. Falcon Observe evaluates and explains whether agents behave as intended. Together, they secure AI operations in the agentic enterprise.

Forward-Looking Statements

This blog contains forward-looking statements, including statements regarding the closing and benefits of the proposed acquisition. These statements involve risks and uncertainties, and actual results may differ materially. There are a number of risks that could cause actual results to differ materially, including the satisfaction of the acquisition’s closing conditions, our ability to integrate Fiddler, and other risks described in the filings we make with the U.S. Securities and Exchange Commission from time to time.