Snowflake and Saviynt enhance governance for autonomous AI agents

Snowflake has released the fifth edition of its ‘Modern Marketing Data Stack’ report, titled ‘Governing the Agentic Enterprise’. The study highlights a shift in artificial intelligence from simple assistance to autonomous action through ‘agentic workflows’. This evolution makes data governance a critical driver of performance and security, as the quality of AI agent decisions depends directly on data reliability, access rules, and control mechanisms.

To address the security risks posed by these autonomous agents, Saviynt and Snowflake are partnering to integrate controls for third-party AI agents. Because traditional identity security models were designed for humans rather than autonomous agents, a security gap exists when agents access sensitive data or call external tools. The new integration uses Saviynt Access Gateway to analyze an agent’s intent, context, and risk level in real time, ensuring agents can only execute actions necessary for their assigned tasks and preventing the abuse of privileges.