With data becoming central to an effective enterprise AI strategy, the need to safeguard sensitive information with robust, enterprise-grade encryption is more important than ever. RelationalAI’s relational knowledge graph fills a critical gap in the enterprise AI infrastructure and together with Snowflake, to build intelligent applications. It extends the Snowflake platform with compound AI capabilities for conceptual modeling, rule based reasoning, graph analytics and prescriptive analytics. At RelationalAI, we understand that trust is built on transparency and control. That's why RelationalAI now supports end-to-end encryption using Tri-Secret Secure (TSS) for customers running the RelationalAI native application on Snowflake’s Snowpark Container Services (SPCS).
Security for Your Most Sensitive Data
One of the key security features that Snowflake customers rely on—especially those on the Business Critical edition and above—is Tri-Secret Secure (TSS) encryption. This ensures that data is always encrypted using customer-managed keys (CMK), meaning organizations retain full control over access. This layered encryption model gives customers full control—at any moment, they can revoke access to their data simply by disabling the key in their own cloud-based key management service (e.g., AWS KMS, Azure Key Vault etc.)
RelationalAI’s native application is designed to run directly within the Snowflake environment, taking full advantage of Snowflake’s industry-leading security infrastructure. For customers already using Snowflake’s TSS is a familiar capability that ensures data at rest is encrypted.
Extending Tri-Secret Secure to RelationalAI
At Snowflake summit 2025, RelationalAI made key announcements with respect to enhanced reasoning capabilities in the native app. With our latest version of the application, RelationalAI Native app has extended this encryption paradigm to its entire footprint, ensuring that:
All data at rest used or stored by the RAI application—whether in internal stages or query caches—is protected by TSS. Customers retain full control over encryption keys, maintaining sovereignty over access and visibility into key usage.
This ensures that all aspects of the RelationalAI workload are protected under the same rigorous standards already in place for Snowflake’s core data storage. This integration means no new security models to learn or configure. Customers benefit from a consistent experience across their Snowflake environment—whether dealing with native Snowflake features or advanced reasoning workloads powered by RelationalAI. This makes it easier for data teams, compliance officers, and security engineers to uphold their organization’s data governance policies without adding operational complexity.
TSS encryption for RelationalAI is supported for the Snowpark Container Services (SPCS) on AWS and Azure.
Closing Thoughts
With RelationalAI’s native integration into Snowflake and its adoption of Tri-Secret Secure encryption, enterprises can confidently build AI-powered applications with a relational knowledge graph grounded on their most sensitive datasets.
To learn more about RelationalAI’s or to get started with our native app on Snowflake, reach out to our team or visit our documentation.