Snowflake rolls out Snowflake Intelligence, Cortex Code updates

Snowflake said it has updated Snowflake Intelligence and Cortex Code in a bid to become a single control plane for developers and business users to move AI agents into production.

The move by Snowflake highlights how multiple enterprise software vendors are looking to become the control plane, context layer and governance platform for agentic AI deployments.

Snowflake said the updates ground agents with data on its platform. Snowflake Intelligence provides a personalized context-aware AI agent grounded in enterprise data to turn insights into action. Cortex Code has a set of AI-powered development tools to move code to production.

According to Snowflake, Snowflake Intelligence will serve "as a personal work agent for business users" and Cortex Code "is expanding as a builder layer for enterprise AI." Snowflake is playing at the intersection of technical and business use cases.

• Snowflake launches Project SnowWork, eyes AI workflows
• Snowflake delivers strong Q4 amid data, AI demand
• Snowflake, OpenAI ink $200 million multi-year partnership
• Snowflake: Cortex Code, Semantic View Autopilot GA
• Snowflake acquires Observe, expands into telemetry data observability

Here's a look at what Snowflake announced for Snowflake Intelligence and Cortex Code.

Snowflake Intelligence is looking to provide a unified experience with the following updates:

• Skills, which will be generally available shortly, can automate routine tasks so business users can describe workflows in natural language and have them executed.
• Model Context Protocol (MCP) connectors that will be available soon to connect Google Workspace, Jira, Salesforce and Slack.
• Deep research with multi-step reasoning to work through and reason over structured and unstructured content.
• Reusable artifacts and personalization over time. Snowflake Intelligence will retain behavior, automate recurring tasks and reuse work.

The catch is that the updates, designed to build on the launch of Project SnowWork, are either generally available in the near future or in public preview.

For Cortex Code, Snowflake said that more than half of its customers are leveraging Cortex Code since it launched in November. Cortex Code gets the following as Snowflake aims to be an AI coding agent for enterprise data.

• Cortex Code expanded its data systems access to AWS Glue, Databricks and Postgres. Snowflake also said Cortex Code will plug into other AI systems through MCP and Agent Communications Protocol.
• A VS Code extension, which is in private preview, and a Claude Code plug-in can connect Cortex Code to multiple integrated development environments.
• Snowflake launched a new Agent Software Development Kit for Cortex Code that supports Python and TypeScript to connect to enterprise apps and workflows.
• Cortex Code Sandboxes in Snowsight will be in private preview for a dedicated cloud environment.