June 2026
A special thank you to everyone who joined us at Snowflake Summit 2026.and Databricks Data + AI Summit. We appreciated the opportunity to connect, exchange ideas, and have meaningful conversations about the future of AI data security and governance.
We are thrilled to announce the enhancements to the current Guardian Agent, our next-generation security agent and a conversational security expert designed to revolutionize how you manage security, compliance, and data governance. With the June release we announce Dremio as a first-class supported data source system.
Feature List
Guardian Agent Overview and Enhancements
RBAC and ABAC support for Dremio
Privilege Reduction
Data Bricks Workspace Management
RBAC Support for Databricks AI Apps
Risk Detection

New Features & Enhancements
Guardian Agent
TrustAI Guardian is an autonomous security layer offering a natural-language interface to identify, prioritize, and remediate data risks for users, service accounts, and AI agents. In this release we have significantly improved the features and performance of the Guardian Agent. TrustAI Guardian Agent is now built as a multi-agent system, not a single copilot or rules engine and provides the following benefits to CISO:
Continuous reasoning: Agents can correlate identity, data sensitivity, behavior, and policy over time.
Autonomy with guardrails: Move from alerts to recommendations to controlled remediation.
Scalability: Match the velocity of AI agents and automated access.
Explainability: Convert low-level telemetry into human-understandable narratives.
Automated Compliance: Stay ahead of evolving regulations with built-in governance mappings, from traditional standards (GDPR, SOC2, NIST CSF) to cutting-edge AI frameworks (ISO 42001 and the EU AI Act).

New Features & Enhancements
Dremio Support
RBAC and ABAC Support for Dremio: We are proud to introduce Dremio as a fully supported, first-class data source. We now deliver the same granular security and policy management you use for Snowflake and Databricks directly to Dremio. Easily build and enforce role-based (RBAC) and attribute-based access controls (ABAC) across your entire data ecosystem without changing your workflow.
Privilege Reduction: Enhanced our security posture by minimizing the required permissions for service principals. Specifically, we have eliminated the 'account admin' requirement for Databricks service principals and the 'manage grant' requirement for Snowflake roles, allowing policies to be enabled at the workspace level.
Databricks Workspace Management: Introduced data source-level workspace selection indicating what metrics are shown across all workspaces.

RBAC Support for Databricks AI Apps: Added role-based access control policies for Databricks Apps, enabling differentiated access for administrators, developers, and end users through group-based permissions integrated with Okta and managed through enhanced backend APIs, Trustlets, and policy management workflows.
Risk Detection: Four new out-of-the-box risk detector rules identify agents inheriting excessive permissions from users, such as running with account admin roles:
Sensitive Data Exposure: Are agents accessing any PII or sensitive data?
Excessive Permissions: What over-privileged permissions are agents deriving from a user’s historical usage?
Data Exfiltration: Are agents performing unauthorized data movement or exfiltration activities?
Compliance Violations: Are any agents breaching jurisdictional data residency regulations?

Each month, we select one critical piece of research or news to help you stay ahead of the curve.
Last updated
Was this helpful?

