Snowflake acquires Natoma to expand AI agent governance capabilities

Snowflake announced its plans to acquire Natoma , a startup focused on securing and managing how AI agents connect to enterprise systems and data through the Model Context Protocol (MCP).

Founded in 2024, Natoma provides tools that help companies control what AI agents can access and do across applications, databases, APIs, and internal systems, with features such as permissions management, policy enforcement, monitoring, and audit trails. The company raised $7 million in seed funding in 2025 from Greylock Partners and Index Ventures.

The technology will be integrated into Snowflake’s AI platform to help customers securely connect AI tools such as Cortex Agents and Cortex Code to SaaS applications, cloud environments, internal databases, Slack, Jira, CRM systems, email, and APIs.

“AI agents are quickly becoming part of how enterprises operate, but intelligence without governance creates risk. Agents don’t just need access to data. They need the right context, permissions and policy guardrails to operate safely inside the enterprise.” – Sridhar Ramaswamy, CEO of Snowflake.

“AI agents will only become enterprise-ready if organizations can govern how they operate across systems, applications and tools.” – Pratyus Patnaik, Co-Founder and CEO of Natoma.

The deal comes as insurers and insurance technology providers continue expanding the use of generative AI and autonomous agents across underwriting, claims, customer service, and internal operations, increasing the need for governance and access controls around enterprise AI systems.