Exabeam has expanded its security operations platform with new capabilities to help enterprises detect, investigate, and mitigate risks associated with AI agents, autonomous workflows, and human-to-agent activity.
The July release broadens Exabeam’s Agent Behavior Analytics, Outcomes Navigator, Nova, Threat Center, Attack Surface Insights, and data collection workflows.
For channel partners, the update reflects a growing opportunity in services around AI governance and SOC modernization.
MSPs and MSSPs are increasingly being asked to help customers secure AI tools already entering enterprise workflows, and Exabeam’s expanded detections and telemetry capabilities could give partners another way to assess AI usage, identify behavioral risks, and build managed detection services around agentic AI.
Exabeam targets risks from agentic AI
AI agents are becoming active enterprise users, capable of accessing systems, invoking tools, interacting with data, and acting on behalf of human employees.
That creates a security challenge for SOC teams because potentially risky activity may still occur through approved applications, valid credentials, and authorized workflows, not to mention lingering risks stemming from misconfigurations, permission issues, and other security threats introduced by agentic AI into organizations.
“Organizations are rapidly moving from AI experimentation to autonomous AI agents operating across the enterprise,” said Pete Harteveld, CEO of Exabeam. “Security teams need visibility not only into human activity, but into how agents behave, interact, and make decisions.”
Exabeam said its Behavior Intelligence model connects behavioral detection, AI-driven investigation, automation, and outcomes-based measurement to identify misuse or compromise across human users, AI agents, and the systems they interact with.
New detections expand AI coverage
The release adds new AI- and agent-focused behavioral detections, bringing Exabeam’s AI detection coverage to 90 in total.
These include indicators tied to:
- suspicious prompt behavior
- unusual tool invocation sequences
- abnormal consumption patterns
- unauthorized configuration changes
- denial-of-wallet activity
- shadow AI usage
- and other signs of misuse.
Exabeam is also expanding visibility across enterprise AI platforms. The company now supports Anthropic Claude alongside OpenAI ChatGPT, Google Gemini, Microsoft Copilot, and GitHub Copilot, giving security teams broader insight into which AI tools are being used and how adoption patterns are changing.
The update also adds coverage aligned to the OWASP Top 10 for Agentic AI through Exabeam Outcomes Navigator. That capability is intended to help customers assess existing detection coverage, identify gaps, and prioritize improvements as agentic AI risk models mature.
Open source telemetry project supports agent observability
Exabeam also introduced Observra, a new open-source project and library that provides developers, security teams, and platform teams with a telemetry layer for AI agents.
Observra captures agent activity, normalizes it into consumable events, enriches it with cost, redaction, deduplication, and risk signals, and routes the information to security operations platforms.
The company positioned Observra alongside Praxen, its previously released open source project for Agent Behavior Verification, as part of a broader agent security lifecycle.
“AI agents introduce a new security challenge because organizations need confidence both before and after deployment,” said Steve Wilson, chief AI officer at Exabeam and co-founder and co-chair of the OWASP Gen AI Security Project.
The release also includes SOC workflow enhancements, such as phishing email ingestion, new cloud collectors, custom REST API context collection, dashboard authoring, global search updates, and expanded integrations with the LogRhythm SIEM ecosystem.