What’s New in New-Scale January 2026: AI Agent Security Is Here

The widespread adoption of AI agents creates a new, unmanaged attack surface. The Exabeam January New-Scale launch directly addresses this risk, providing a layered approach to secure your AI workforce from onboarding to automated threat response.

An Ungoverned AI Workforce Is Already Here

Nearly 80% of enterprises deploying AI agents, creating a new digital workforce with access to sensitive data and critical systems. This rapid, often ungoverned adoption introduces significant risk, as a single malicious prompt can turn a helpful AI tool into an insider threat operating at machine speed. According to Gartner, AI-enhanced malicious attacks have been the top emerging risk for enterprises for three consecutive quarters.

The critical question is no longer if you will adopt AI, but how you will secure it. An ad-hoc measures are insufficient. Securing this new workforce requires a dedicated, end-to-end security framework.

Step 1: Onboard AI Agents and Give Them Secure Access to Data

Effective AI agents require access to data. To provide this access securely, the principle of least privilege is critical. Our January launch makes it easier to onboard AI agents securely. New prebuilt log collectors for Gemini Enterprise Security speed up ingestion, while updates to the Exabeam Common Information Model (CIM) enrich data with fields specific to AI use and agent activity.

The new Exabeam Model Context Protocol (MCP) Server acts as a secure data gateway, allowing external AI agents to consume contextual data from the platform without direct, risky integrations.

Step 2: Real-Time Security with Behavioral Analytics for AI Agents

Once an agent is active, the challenge becomes monitoring its behavior. Consider these real-world scenarios:

• An agent designed to generate quarterly financial summaries is tricked by a malicious prompt. It begins pulling all files from the finance department and sending them to an unauthorized external API endpoint.
• A helpdesk bot, normally used to reset passwords during business hours, is compromised. At 2 AM on a Saturday, it begins systematically deleting critical security logs to cover an attacker’s tracks.

These scenarios represent the new face of insider risk. They are also invisible to legacy SIEM or XDR tools that can only find the threats they have been told to look for.

The only way to secure this new attack surface is to understand its behavior. New-Scale Analytics, our flagship UEBA platform, now supports Agent Behavior Analytics (ABA)—an industry first that applies our proven behavioral analytics directly to this non-human workforce.  With an initial offering of AI agent behavioral detections, we can instantly detect suspicious deviations that indicate a compromise or agent misuse. When an agent starts exploring systems outside its function or accessing unusual volumes of sensitive data, we spot it. When a threat is detected, our platform automatically provides a forensic timeline of every action, eliminating hours of manual log analysis.

Step 3: From Agent Security to Strategic Oversight

Stopping individual threats is only the first step. Security leaders must also manage overall risk, demonstrate compliance, and communicate posture to the business. The latest updates to Outcomes Navigator deliver this strategic oversight.

A new Agentic AI Security use case provides a dedicated, board-ready dashboard to translate complex agent activity into a clear view of your AI risk, helping you answer critical questions like, “How much of our AI activity is being monitored?” and “Where are our biggest security gaps?”

Additionally, the January release includes a Compliance Posture Assessment. For the first time, you can get a quantitative score and a clear view of your coverage for top compliance frameworks like NIST and GDPR. The Exabeam Nova Advisor Agent generates strategic reports with prescriptive recommendations to improve your compliance score.

Platform-Wide Enhancements in the January Launch

The January launch extends beyond AI security with platform-wide enhancements. We have improved the analyst experience with redesigned dashboards and a more intuitive timeline to make investigations and response faster and more efficient. We also strengthened the core of the New-Scale Platform by expanding our data coverage with new Cloud Collectors, enhancing APIs for deeper integrations, and adding new security controls like Universal Logout for Okta.

Building the Accountable Hybrid Team

The risks of unmanaged AI are immediate, automated, and too fast for manual investigation. A modern security strategy requires three layers of AI agent security: a secure data foundation, real-time behavioral monitoring, and strategic risk management.

The New-Scale Platform delivers this with the MCP Server for secure onboarding, ABA for threat detection, and Outcomes Navigator for program management. With these capabilities, you can build an accountable hybrid team where humans lead strategy, AI accelerates work, and the entire system is secure, compliant, and ready for the future.

Ready to Secure Your AI Workforce?

Register for our upcoming webinar, “What’s New” with the Exabeam New-Scale Platform—A blueprint for Securing AI Agents,” to see these new capabilities in action.