Snowflake announced a series of major innovations to its platform designed to help organisations deploy agentic AI at scale and deliver trusted, actionable insights faster than ever before.
These enhancements include the general availability of Snowflake Intelligence, an enterprise intelligence agent that enables every user to ask complex business questions in natural language and instantly access insights. Together with advancements in Snowflake Openflow and Snowflake Horizon Catalog, these updates allow enterprises to connect and harness all their data — structured, semi-structured, and unstructured — from disparate sources and catalogues to power next-generation AI applications.
All of this is delivered within Snowflake’s secure, governed, and interoperable environment, free from vendor lock-in.
Snowflake also introduced a new suite of AI-native developer tools that make it easier for teams to build, test, and deploy production-ready AI applications with improved reliability, reduced overhead, and lower total cost of ownership — all within a single, governed platform.
“For more than a decade, Snowflake has been a cornerstone of global data strategies,” said Christian Kleinerman, EVP of Product at Snowflake. “Our next evolution is about bringing AI to that data, helping every customer unlock intelligence that is uniquely their own. These latest enhancements democratise the power of AI so every employee can make faster, smarter decisions — fundamentally transforming how our customers innovate.”
Snowflake Intelligence: All Your Knowledge, One Trusted Enterprise Agent
Snowflake Intelligence is now generally available to over 12,000 Snowflake customers worldwide. With a single query, users can conduct in-depth analysis, surface insights, and resolve complex business challenges — moving beyond the what to uncover the why.
Built with trust, governance, and scalability at its core, Snowflake Intelligence ensures employees can confidently interact with enterprise data through natural language — while keeping sensitive information secure. This empowers data-driven decision-making and fosters a new culture of intelligence across organisations.
Over the past three months, more than 1,000 customers — including Cisco, Toyota Motor Europe, TS Imagine, and the USA Bobsled/Skeleton Team — have deployed 15,000+ AI agents using Snowflake Intelligence.“Snowflake Intelligence has transformed our development timeline, reducing agent deployment from months to weeks,” said Thierry Martin, Head of Data and AI, Toyota Motor Europe. “It allows us to focus on what drives value — building strong business context and robust semantic models — while bringing secure, compliant solutions to market faster.”
Powered by advanced models from providers such as Anthropic, Snowflake Intelligence turns complex queries into conversational insights, helping to democratise access to data and AI.
New innovations from Snowflake’s AI Research Team have made text-to-SQL queries up to three times faster, while the novel Agent GPA (Goal, Plan, Action) framework detects up to 95% of query errors, achieving near-human accuracy in evaluation.
Delivering the Enterprise Lakehouse: Enhanced Open Data Access for Agentic AI
Snowflake has also announced major enhancements to Snowflake Horizon Catalog and Snowflake Openflow (both now generally available), enabling enterprises to connect disparate data sources and catalogues to power more robust and accurate AI-driven insights.
New innovations to Horizon Catalog introduce a unified security and governance framework that connects and secures data across every region, cloud, and format — all interoperable and without vendor lock-in. Meanwhile, Openflow allows enterprise users to automate data integration from virtually any source, maintaining centralised control across the enterprise lakehouse.
Additional updates include:
Open APIs from Apache Polaris™ (Incubating) and Apache Iceberg™ REST Catalog integrated into Horizon Catalog, offering customers centralised governance, security, and interoperable access management across open table formats.
Interactive Tables and Warehouses (private preview): redefine how enterprises build and power AI agents by enabling real-time insights and experiences.
Near real-time streaming analytics (private preview soon): allows organisations to act on live data within seconds, supporting use cases such as fraud detection, personalisation, and IoT monitoring.
Expanded Oracle integration (private preview): continuous data streaming into the Snowflake AI Data Cloud via Openflow-based change data capture.
Snowflake Postgres (public preview soon): a fully managed service following Snowflake’s acquisition of Crunchy Data, bringing Postgres directly onto the Snowflake platform.
pg_lake (now generally available): open-source Postgres extensions enabling developers to integrate Postgres with a modern lakehouse architecture.
Business Continuity and Disaster Recovery (public preview): enhanced protection for managed Iceberg tables to ensure enterprise data resilience.
Supercharging Agentic AI development: New developer tools
Snowflake unveiled a comprehensive suite of developer tools to accelerate and streamline enterprise-grade AI app development:
Cortex Code (private preview): an AI assistant within the Snowflake UI that allows users to interact with their entire Snowflake environment using natural language, optimise queries, and gain cost-saving recommendations.
Snowflake Cortex AISQL (now generally available): enables developers to build scalable AI pipelines directly within Dynamic Tables, creating inference pipelines using simple SQL queries.
AI Redact (public preview soon): automatically detects and removes sensitive data from unstructured datasets to ensure privacy and compliance.
Workspaces (now generally available): a unified development environment that enhances collaboration with direct Git and VS Code integrations, allowing teams to co-develop securely.
dbt Projects on Snowflake (now generally available): allows engineers to build, test, and deploy dbt projects natively within Snowflake.
Snowpark Connect for Apache Spark™ (now generally available): enables organisations to run existing Spark code securely within Snowflake’s environment.