Exabeam launches AI Agent Security: governs agented tooling

Exabeam added AI Agent Security in its New-Scale January 2026 release. The feature sets AI-driven security workflows to onboard, monitor, and score agent behavior. The vendor announced the update on Jan 6, 2026 and frames it as an extension of UEBA to non-human agents.

The release includes Agent Behavior Analytics (ABA) that applies UEBA models to agent activity. It also adds a Model Context Protocol (MCP) Server as a secure data gateway for external agents. Outcomes Navigator gains a Compliance Posture Assessment. Exabeam shipped prebuilt log collectors for Gemini Enterprise to speed telemetry ingestion. The vendor documents these features in its blog and related coverage.

The update targets a governance gap. Exabeam cites data showing roughly 80% of enterprises already use AI agents (PwC). The platform gives unified timelines, detections for agent drift or misuse, and a quantitative posture score to prioritize remediation. Security teams can plug these tools into existing SIEM and UEBA workflows and test integrations with identity and data controls.

Why It Matters

• Centralizes agent visibility: ABA applies UEBA baselines to non-human agents, enabling detection of anomalous actions such as data exfiltration or unusual access patterns.
• Secure onboarding: the MCP Server provides a contextual data gateway that reduces risky direct integrations and enforces secure agent connections.
• Operational posture: the Compliance Posture Assessment converts agent telemetry into a board-ready score and prioritized remediation tasks.
• Faster integration: prebuilt Gemini Enterprise collectors and platform APIs shorten telemetry onboarding for organizations using Google’s agent stack.