Snowflake Summit Day 1: AI Agents for Clinical Systems

Healthcare organisations could benefit from new tools designed to help manage data governance for AI agents operating across clinical and administrative systems.

According to Snowflake's The ROI of gen AI and Agents research, 96% of organisations face challenges with scaling AI across the enterprise.

Addressing fragmented clinical systems

Snowflake's acquisition of Natoma aims to address these challenges through an enterprise Model Context Protocol platform for AI agents.

"The future of AI is about turning fragmented systems into actionable insight," says Sridhar Ramaswamy, CEO of Snowflake, in the keynote for Snowflake Summit 26, taking place at the Moscone Center in San Francisco, California, on 1-4 June.

"Intelligence without governance creates risk," Sridhar added. "Agents don't just need access to data. They need the right context, permissions and policy guardrails to operate safely inside the enterprise."

Healthcare providers operate across multiple disconnected platforms including electronic health records, laboratory information systems, radiology imaging platforms and billing software. This acquisition could allow AI agents to access patient data, treatment protocols and operational information across these systems whilst maintaining HIPAA compliance and audit trails.

Natoma provides control and governance for these connections, helping healthcare organisations manage how AI agents discover, access and act across systems with identity-aware authorisation, policies and auditability.

Connecting clinical and operational data

Data held in Snowflake can be correlated with contextual information from clinical communication platforms, EHR systems, care management tools and internal APIs. This could mean that AI agents supporting clinical decision-making or administrative tasks can access information from multiple sources without requiring manual data aggregation by staff.

"Agents behave differently than human users. They can explore paths, call APIs and attempt workflows in ways that require clear boundaries and oversight," explains Mayank Upadhyay, Chief Security and Trust Officer and VP of Engineering at Snowflake.

"Enterprises need a way to give their people the magic of agentic productivity whilst maintaining a single point of control over what those agents can access and do. That's why we announced our intent to acquire Natoma."

For healthcare settings, this could apply to scenarios where AI agents retrieve patient histories from EHR systems, cross-reference lab results, check medication interactions from pharmacy databases and access insurance eligibility information from billing systems whilst maintaining role-based access controls.

The acquisition aims to simplify deployment by connecting AI intelligence with healthcare security requirements. Healthcare organisations typically operate dozens of specialised systems. Pre-configured connections could reduce the technical work required to enable AI agents across these environments.

Security enabling clinical workflows

The technology aims to make complex backend governance feel effortless for clinical users, transforming security and compliance teams from restrictive gatekeepers into workflow enablers.

"The end state, once Natoma is fully integrated, should feel invisible to the user," says Mayank. "An admin flips a switch and their people don't have to do anything differently. They simply notice that the AI tools they're already using can suddenly access far more information on their behalf, all within a governed environment.

"That's security becoming an enabler rather than a blocker. That's the standard we're building towards and Natoma is the next major step."

For physicians, nurses and other clinical staff, this could mean AI tools can retrieve relevant patient information without requiring users to log into multiple systems. Compliance and security teams could monitor and audit all AI agent activity from a single control panel.

Snowflake has deployed Natoma across its own organisation with immediate results. Healthcare applications could include AI agents that summarise patient messages from portal systems, retrieve relevant clinical guidelines from knowledge bases and surface relevant case notes from disparate sources.

Natoma's team brings expertise in Model Context Protocol (MCP), gateway infrastructure, identity governance and privileged access management. This expertise could help healthcare organisations scale automation efforts whilst maintaining compliance with HIPAA, GDPR and other healthcare data protection requirements.

"AI agents will only become enterprise-ready if organisations can govern how they operate across systems, applications and tools," says Pratyus Patnaik, Co-Founder and CEO of Natoma. "Together with Snowflake, we're building the governance and connectivity layer that enables enterprises to securely operationalise AI at scale."

Natoma's capabilities will be integrated into Snowflake's AI Data Cloud and available to customers soon.