Snowflake announces updates for Snowflake Intelligence and Cortex Code. Together, the tools are designed to function as a control layer for the “Agentic Enterprise.” Snowflake Intelligence gains personalization, MCP integrations, and deep research capabilities. Cortex Code expands to support more data systems and development environments.
Snowflake today announced significant updates for Snowflake Intelligence and Cortex Code. Together, they form what the company calls the “Control Plane for the Agentic Enterprise.” This is a central layer that enables organizations to connect data, tools, and AI agents.
Snowflake Intelligence as a personal work agent
Snowflake Intelligence is evolving into a personal work agent for business users. The tool learns from individual user behavior and personalizes responses based on previous work. New features include Skills, which allow users to describe workflows in plain language. Think of automatically creating presentations or sending follow-ups. Snowflake then executes those tasks independently.
New MCP connectors link Snowflake Intelligence directly to Gmail, Google Calendar, Jira, Salesforce, and Slack. In addition, there is an iOS app that allows users to request analyses on the go. With the “deep research” feature, Snowflake Intelligence answers complex questions via an agentic architecture that combines both structured and unstructured data.
“Snowflake gives customers one place to bring their data together, connect the systems they rely on, and turn AI into something that actually helps teams get work done,” says Baris Gultekin, VP of AI at Snowflake.
Cortex Code Now Available Across a Broader Data Stack
Cortex Code was launched in February 2026 as an AI coding agent for the enterprise data stack. Since then, more than 50 percent of Snowflake customers have been actively using the tool. Now, Cortex Code is expanding its support to external data systems, including AWS Glue, Databricks, and Postgres.
Additionally, developers can access Cortex Code via a VS Code extension and a Claude Code plugin, allowing them to work from their own development environments. A new Agent SDK with Python and TypeScript support enables the integration of Cortex Code into custom applications. In Snowsight, users can run complete workflows through Cortex Code Sandboxes without a local installation.
Cortex Code also connects to other AI systems via MCP and the Agent Communication Protocol (ACP). This allows builders to access the tool from existing AI agents and workflows.