Identity security firm Saviynt has launched a new integration with Snowflake, the AI Data Cloud company, aimed at reducing identity risk as enterprises increasingly grant third-party AI agents access to business systems. The announcement, dated from Singapore, comes as adoption of agentic AI accelerates across industries in the region and beyond, raising fresh governance questions for security teams.

The integration allows organisations using Snowflake’s security ecosystem to securely connect, govern and audit AI agents operating across business platforms, addressing a gap that traditional identity frameworks — built around human users — were not designed to close.

Closing the gap traditional identity models were not built for

As AI agents take on more autonomous decision-making, accessing sensitive data and invoking external tools, security teams face governance challenges that legacy identity and access management tools struggle to address. Saviynt’s Access Gateway is designed to evaluate an agent’s intent, context, permissions and risk at the point of access, rather than granting standing privileges upfront.

“Traditional identity security models were built for human users, not AI agents, creating a critical gap in enterprise security. Our integration with Snowflake addresses this challenge through Saviynt’s Access Gateway, which evaluates agent intent, context, permissions, and risk at runtime,” said Akhil Kapoor, Senior Vice President, Business Development at Saviynt.

Key elements of the integration include:

• Runtime evaluation of AI agent intent, context, permissions and risk before granting access
• Governance and audit capabilities for third-party agents operating across business systems and platforms
• Alignment with Snowflake’s existing security ecosystem to support production-ready AI deployments

A broader industry push toward agent interoperability

Snowflake framed the integration as part of a wider industry shift toward securing what it calls the “agentic enterprise” through cross-vendor collaboration, rather than isolated point solutions.

“Securing the agentic enterprise requires an entirely new model of trust, visibility, and control. Through our integration with Saviynt, we are establishing a secure foundation for agent interoperability,” said Mayank Upadhyay, Chief Security and Trust Officer, Snowflake.

The companies said the integration marks a step toward establishing the trusted foundation needed to run next-generation AI agents safely in production environments, as enterprises move from experimenting with AI copilots toward deploying autonomous, task-executing agents at scale.