Exabeam Enhances AI Agent Analytics for Cybersecurity Monitoring

• •Exabeam's analytics now support ChatGPT and Microsoft Copilot.
• •The update enhances monitoring of AI agent interactions and behaviors.
• •Detection capabilities for prompt injection and model manipulation have significantly expanded.

Exabeam has expanded its Agent Behaviour Analytics product to include support for OpenAI ChatGPT and Microsoft Copilot, enhancing visibility into interactions between employees and AI agents within enterprise systems. This update aims to provide security teams with telemetry data to improve threat detection, investigation, and response workflows. As generative AI assistants become more prevalent, cybersecurity teams are encountering new challenges regarding oversight and misuse of these tools. Steve Wilson, Exabeam's chief AI and product officer, emphasized that AI agents are evolving into autonomous digital workers, which complicates security monitoring due to their legitimate-looking activities when compromised. The update includes expanded detection capabilities for prompt injection and model manipulation, as well as lifecycle monitoring for AI agents. Exabeam's detection library has increased fivefold, and the product aligns with the OWASP Top 10 for Agentic AI, establishing a benchmark for security governance in this emerging field.