New behavior analytics, AI agent detections, and open-source Observra help security teams monitor, investigate, and govern autonomous AI agents across enterprise environments.
Exabeam has unveiled a major expansion of its Behavior Intelligence platform, introducing new capabilities designed to help enterprises secure the growing use of autonomous AI agents without slowing innovation.
The latest release enhances Exabeam’s Agent Behavior Analytics (ABA), Exabeam Nova, Outcomes Navigator, Threat Center, Attack Surface Insights, and search and data collection workflows, enabling security teams to detect, investigate, and mitigate risks arising from AI agents, autonomous workflows, and human-to-agent interactions.
As enterprises increasingly deploy AI agents to automate business processes, traditional security controls often struggle to identify malicious or risky behavior because AI agents typically operate using legitimate credentials and approved applications. Exabeam’s Behavior Intelligence approach addresses this challenge by combining behavioral analytics, AI-driven investigations, automation, and outcomes-based measurement to detect suspicious activities over time.
“Security teams need visibility not only into human activity, but into how agents behave, interact, and make decisions.”
— Pete Harteveld, CEO, Exabeam
Pete Harteveld, CEO of Exabeam, said organizations are rapidly transitioning from AI experimentation to enterprise-wide deployment of autonomous AI agents, making visibility into both human and machine behavior essential for effective cybersecurity.
The latest release doubles Exabeam’s AI-focused behavioral detection capabilities to 90 use cases, covering threats such as suspicious prompt activity, abnormal tool usage, unauthorized autonomous actions, shadow AI, denial-of-wallet attacks, configuration tampering, and unusual AI consumption patterns.
Exabeam has also expanded visibility across leading AI platforms by adding support for Anthropic Claude alongside OpenAI ChatGPT, Google Gemini, Microsoft Copilot, and GitHub Copilot, enabling security teams to monitor AI adoption and usage across enterprise environments.
To strengthen AI risk management, the company’s Outcomes Navigator now aligns detection coverage with the OWASP Top 10 for Agentic AI, allowing organizations to identify protection gaps and prioritize security investments against emerging AI threats.
The release further enhances security operations through Exabeam Nova Rules Creator, which enables analysts to create and refine detection rules using natural language and convert Sigma rules automatically. A new Related Cases capability helps SOC analysts identify connected incidents faster by correlating shared entities such as IP addresses and hosts.
Additional platform improvements include phishing email ingestion, enhanced attack surface insights, expanded cloud data collectors, REST API integrations, improved dashboard creation, global search enhancements, and advanced reporting capabilities.
Alongside the platform enhancements, Exabeam introduced Observra, a new open-source telemetry framework that captures, normalizes, and enriches AI agent activity across multiple AI frameworks before forwarding the data to security platforms. The project complements Exabeam’s previously launched Praxen initiative, which focuses on verifying AI agent configurations and governance before deployment.
According to Steve Wilson, Chief AI Officer at Exabeam and Co-Chair of the OWASP Gen AI Security Project, organizations must verify AI agents before deployment while continuously monitoring their behavior throughout their operational lifecycle to ensure they remain secure and compliant.
The company also announced expanded LogRhythm SIEM integrations across Microsoft, cloud, identity, email, and security technologies to provide broader visibility across modern enterprise attack surfaces while simplifying security operations.
