What's new

Explore our latest product enhancements, release notes, acquisitions, and more.

Apr 2026

Platform (Atlas) AI Security

Browser-based Pentests

Varonis Atlas now supports browser-based pentesting, enabling security teams to assess AI applications that are accessed and exercised through the browser. This expands testing coverage beyond direct API and endpoint interactions.
Platform (Atlas) AI Security

Multi-modal guardrails

Varonis Atlas customers can now use multi-modal guardrails to extend runtime protection beyond text-only interactions. This allows guardrails to evaluate and enforce policy across additional content types handled by supported AI applications.
Platform (Atlas) AI Security

AWS Assumed Role authentication in the Gateway

Varonis Atlas now supports AWS Assumed Role authentication in the Gateway. This enables customers to route requests through the Gateway while using AWS role assumption for secure downstream access, improving support for Bedrock and other AWS-based integrations.
Platform (Atlas) AI Security

Pentest Strategies

Varonis Atlas customers can now use Pentest Strategies to improve how attack approaches are organized and executed during AI red team assessments. This creates a more structured foundation for test coverage and makes it easier to expand and compare different attack methods over time.
Platform (Atlas) AI Security

AI Investigation anomaly detectors

Varonis Atlas now supports anomaly detectors in AI Investigation to help identify unusual behavior patterns and outliers in runtime activity. These detectors improve visibility into potentially risky or unexpected changes in AI system behavior.
Platform (Atlas) AI Security

Discovery for AWS Bedrock AgentCore

Varonis Atlas now supports discovery for AWS Bedrock AgentCore. Customers can now identify and inventory AWS Bedrock AgentCore resources as part of their broader AI asset discovery workflows.
Platform (Atlas) AI Security

Runtime policy automations

Varonis Atlas enables automated responses tied to runtime policy outcomes with Runtime Policy Automations. This helps security teams operationalize runtime detections more efficiently and reduce manual follow-up for common actions.