Snowflake announced updates to its AI Data Cloud on April 21, 2026, focusing on two main pillars: Snowflake Intelligence and Cortex Code. These tools help organisations move beyond simple chatbots toward autonomous AI agents that perform tasks across different software environments.
Snowflake Intelligence for business users
Snowflake Intelligence functions as a personal agent for employees. It connects to company data to answer questions and execute specific tasks. Unlike general AI assistants, this system maintains a connection to the company’s internal security and governance rules.
Key features include:
- Task automation: Users can describe a workflow, such as drafting a report or analyzing data trends, and the agent performs the steps.
- Third-party connections: New connectors link the platform to tools like Gmail, Salesforce, Slack, and Jira.
- Deep research: The system conducts multi-step analysis, providing reports with citations from structured and unstructured data.
- Mobile access: A new iOS app allows users to interact with their data while away from their desks.
Cortex Code for developers
Cortex Code serves as the technical layer for building AI applications. More than half of Snowflake’s customers currently use these tools to write and deploy code. The latest updates expand where and how developers can work.
Builders can now use Cortex Code within external data systems like AWS Glue, Databricks, and Postgres. Snowflake also introduced a VS Code extension and a Claude Code plugin, allowing developers to stay within their preferred coding environments. A new Software Development Kit (SDK) for Python and TypeScript helps teams build these AI capabilities directly into their own custom applications.
Real-world applications
Several companies shared data regarding their use of these tools:
- Telenav: The company processes 20 terabytes of data monthly. They reported that converting raw data into usable insights now takes minutes rather than weeks.
- Wolfspeed: The manufacturer uses AI agents to monitor equipment performance and identify potential process issues before they cause delays.
- United Rentals: Staff at over 1,600 locations use natural language queries to check fleet availability and branch performance without needing a data analyst.
- Capita: This service provider uses the platform to manage operational data for public and private sector contact centers.
System governance and security
Snowflake positions these updates as a control plane. This means the platform centralises how a company manages its AI models and data access. By keeping the AI agents within the Snowflake environment, businesses can track how data is used and ensure that AI actions comply with corporate policies.
Over 9,100 customers currently use Snowflake’s AI products weekly. The company continues to gather feedback through Project SnowWork to refine how these agents learn from user behavior over time.