As enterprises move from dashboards to autonomous agents on structured data, Trust3 AI centralizes access governance once and applies it natively across Unity Catalog, AWS Lake Formation, Snowflake, and all query engines.
Trust3 AI announced the latest release of its centralized data access governance platform at Databricks Data + AI Summit, expanding support for federated catalog governance and multi-engine policy enforcement as enterprises prepare their data assets for agent AI.
At the 2026 Summit, agent AI on structured data will be at the center of the conversation. This change increases risks to governance. When autonomous agents (not just analysts) query the lakehouse, all access decisions must be correct, consistent, and auditable across each catalog and engine that the agents can reach. Most organizations still manage policies on each system individually, which is not scalable and leaves gaps the moment a new engine or agent is added.
Trust3 AI addresses this issue with a single policy administration point (PAP) that delegates enforcement to native catalogs and engines. So one set of policies is created centrally and applied everywhere it needs to exist.
Centralized policy, native enforcement, federated catalog
Large companies are increasingly operating multiple catalogs. Trust3 AI allows organizations to standardize on a single point of policy management while delegating enforcement to native systems like Unity Catalog and AWS Lake Formation. This also includes a new pattern where one catalog acts as the primary and another catalog is federated under it. A Fortune 500 financial software company uses this model to perform fine-grained access control (FGAC) across Lake Formation and Unity Catalog at enterprise scale from a single management layer. At the FGAC level, it is not practical to operate on a catalog-by-catalog basis.
Governance transmitted to every engine
Modern lakehouses built on open formats like Apache Iceberg are regularly queried by multiple engines. Access policies do not automatically track data across these engines, and that very gap is the source of a leak. Trust3 AI propagates a single set of policies across engines such as Databricks/Unity Catalog, Snowflake, AWS Lake Formation, Dremio, Spark, and EMR. Cloud enterprise application providers leverage Trust3 AI to enforce one consistent policy across Iceberg Lakehouse served by multiple query engines, rather than reimplementing rules for each.
Also read: AiThority interview with Matej Bukovinski, Chief Technology Officer at Nutrient
Dynamic policies to eliminate policy sprawl
Static role-based policies grow uncontrollably at scale. Trust3 AI’s Attribute-Based Access Control (ABAC) allows teams to dynamically express intent, reducing the number of policies they manage by orders of magnitude. A global advertising and media network replaced approximately 2,000 catalog policies with approximately 20 dynamic policies. Fortune 50 media and communications companies reported comparable savings of up to 100x while applying the same enforcement to 8+ connectors on both Unity Catalog-enabled and non-Unity Catalog-enabled clusters, including Databricks Spark and Spark on Kubernetes.
Covering new platforms with day zero
Policies exist in one place, so onboarding a new platform immediately inherits existing governance. When the same ad and media network added Snowflake, the central management cascaded the same enforcement to Snowflake starting from day zero. No rebuild required.
Beyond the limitations of native catalogs: data products and purpose-based access
For organizations facing a proliferation of roles in cloud warehouses, Trust3 AI introduces a governed data product with multiple subscriptions and purpose-based access control at the governance layer, rather than individual workarounds within each native catalog. A leading healthcare analytics organization uses Trust3 AI’s data products to manage entitlement and purpose-based access that could not be maintained with native warehouse capabilities alone. Other companies extend this to frictionless, end-to-end access provisioning that eliminates manual steps and incorporates data products defined in tools like Microsoft Purview.
Centralized governance across hybrid real estate
Trust3 AI customers operate a single point of policy management across a heterogeneous stack (a combination spanning Unity Catalog, Starburst, Snowflake, SQL Server, Lake Formation, and EMR) and combine access controls with features like format-preserving encryption to protect and make sensitive data available, with reported productivity gains of up to 10x.
“Lakehouse was supposed to make data accessible. Lakehouse didn’t have a way to consistently manage that access across all the engines and catalogs in the stack. That’s exactly why we built Trust3 AI to solve that. And with agent AI reaching structured data, that need has never been more urgent.” said Don Bosco Durai, co-founder and CTO of Trust3 AI.
At Databricks Data + AI Summit 2026
Trust3 AI will be participating in the Databricks Data + AI Summit June 15-18 in San Francisco. Stop by the coffee truck at Third and Howard Monday through Wednesday from 9 a.m. to 5 p.m. Meet the team to learn more about centralized policy management, federated catalog governance, and multi-engine enforcement.
