Defining AI Agent Governance in the Enterprise
AI agent governance is the set of controls, policies, and monitoring tools that determine how autonomous AI agents access data, use applications, and take actions across enterprise systems under strict compliance, security, and audit rules. Snowflake’s planned acquisition of Natoma puts this discipline at the center of its AI strategy. Instead of treating agents as isolated copilots, Snowflake talks about an “agentic control plane” that coordinates permissions and behavior wherever work happens. Natoma brings a gateway that manages AI agent permissions across enterprise applications so that tasks such as sending emails, opening Jira tickets, or summarizing Slack conversations occur inside a governed environment. This focus on controlled autonomy signals that the next phase of enterprise AI will be judged less by how clever models sound and more by how safely they act inside cloud data platforms and everyday tools.
Natoma and the ‘Rogue Agents’ Problem
Snowflake’s Natoma acquisition targets a growing concern: autonomous AI agents acting as “rogue agents” when they interact with critical business systems. Natoma’s platform sits between AI agents and tools using the Model Context Protocol (MCP), enforcing identity checks, access policies, and detailed audit trails on every tool call. That means each action is tied to a specific user, permission set, and policy decision, reducing the risk of shadow AI and accidental data exposure. Natoma’s founders say they started the company on the belief that agents would only reach production if organizations could trust and control how those agents access data, use tools, and take action. By folding Natoma into Snowflake Intelligence and Cortex Code, Snowflake aims to make its AI interface a single, governed hub for tasks that span email, calendars, tickets, and internal APIs, rather than a loosely monitored collection of bots.
Sixth Snowflake Acquisition and a Shift in Enterprise Priorities
Natoma is Snowflake’s sixth acquisition announcement since mid-2025, underscoring an aggressive expansion strategy around AI agent governance and observability. Earlier deals included PostgreSQL provider Crunchy Data, migration specialist Datometry, data discovery platform Select Star, observability firm Observe, and AI-powered pipeline planner TensorStax. Together, these moves show Snowflake building a stack that runs from core databases to agentic AI control. Where the first wave of generative AI focused on model performance, this sequence of acquisitions marks a pivot toward enterprise AI compliance, permissions, and monitoring. Snowflake cites internal research that 96 percent of organizations still face significant obstacles scaling AI across the enterprise, which aligns with buyers now prioritizing auditability, identity management, and integration with existing workflows. In this context, Natoma is less a niche startup purchase and more a signal that controlling AI behavior is becoming table stakes for cloud data platforms.
AWS Partnership and the Rise of the Agentic Enterprise
Announcing the Natoma deal alongside an expanded AWS partnership shows governance and infrastructure advancing together. Snowflake has committed to a five-year, USD 6 billion (approx. RM27.6 billion) agreement centered on Graviton-powered compute and AI infrastructure, while also outlining a broader multi-year collaboration around generative and agentic AI workloads. One quotable takeaway is Snowflake’s statement that “96% of organizations still face significant barriers scaling AI across the enterprise.” The AWS relationship focuses on keeping AI systems close to governed enterprise data rather than exporting sensitive datasets into external services. Customers will be able to connect Cortex Agents, Snowflake Intelligence, and Cortex Code to SaaS apps, cloud infrastructure, and on-prem systems while staying within Snowflake’s governance layer. This pairing of AI agent governance with large-scale cloud infrastructure suggests that the “agentic enterprise” will be built where data resides—inside secure cloud data platforms with strong control planes.