Snowflake is introducing product innovations at the Build developer conference to deploy agentic AI at scale. The general availability of Snowflake Intelligence is the centerpiece, an enterprise intelligence agent that answers complex questions in natural language. In addition, tools for developers are being expanded to build AI applications faster.
Snowflake Intelligence enables in-depth research and the proposal of solutions to complex business problems with a single question. Users can securely access and analyze data using natural language, while confidential information remains protected.
AI models from partners, including Anthropic power the technology. New innovations from Snowflake’s AI Research Team make the solution up to three times faster for text-to-SQL queries. The Agent GPA (Goal, Plan, Action) Framework captures up to 95 percent of errors in tests with standard datasets.
Developer tools for agentic AI
Snowflake has announced new developer tools that help organizations build production-ready AI applications faster. Developers can streamline data workflows with Cortex Code, an AI assistant within the Snowflake interface that allows users to control their entire environment using natural language.
With enhancements to Snowflake Cortex AISQL, developers can build scalable AI pipelines within Snowflake Dynamic Tables using declarative SQL queries. AI Redact, coming soon in public preview, discovers and masks sensitive data in unstructured data.
Snowflake’s Workspaces development environment eliminates fragmented data workflows and provides a unified editor for creating and managing code. The environment has been expanded with direct Git integration and VS Code integration, allowing users to work from their favorite IDE.
With dbt Projects on Snowflake, organizations can build, test, and deploy their dbt projects directly within their Snowflake environment. In addition, Snowpark Connect for Apache Spark helps organizations run existing Apache Spark code on the Snowflake engine.
Enterprise lakehouse with more open data access
Snowflake announces enhancements to Snowflake Horizon Catalog and Snowflake Openflow to enable companies to connect data from diverse sources. New innovations within Horizon Catalog add context for AI and provide a unified security and governance framework.
By integrating open APIs from Apache Polaris (Incubating) and Apache Iceberg REST Catalog directly into Horizon Catalog, Snowflake offers an enterprise lakehouse that centralizes governance and security across data in open table formats. The company promises full interoperability without vendor lock-in.
With Interactive Tables and Warehouses, now in private preview, organizations can convert data into immediate insights and near-real-time experiences. Near real-time streaming analytics, coming soon in private preview, enables organizations to respond to live data within seconds.
Snowflake is further expanding integration options through a partnership with Oracle. This enables near real-time change data capture based on Openflow to stream continuous transactional updates to the Snowflake AI Data Cloud.
Following the acquisition of Crunchy Data, Snowflake is also introducing Snowflake Postgres, a fully managed service that brings the popular database to the platform. The solution will be available in public preview soon.