Snowflake, the AI Data Cloud company, announced a series of major advancements to its artificial intelligence portfolio on April 23, 2026, specifically targeting the development and deployment of autonomous AI agents. These updates center on Snowflake Intelligence and Snowflake Cortex Code, two core components of the company’s strategy to transition from passive data storage to active, model-driven business execution. CEO Sridhar Ramaswamy stated that these enhancements represent a fundamental shift in how enterprises interact with their data, moving beyond conversational interfaces toward systems that can independently perform multi-step tasks across disparate software environments.
The expansion of Snowflake Intelligence introduces a new framework for actionable agents. This system allows organizations to connect their governed data within Snowflake to external enterprise applications such as Salesforce, SAP, and ServiceNow. By utilizing these connections, AI agents can now execute transactions, update records, and trigger workflows based on real-time data analysis. For example, a supply chain agent can monitor inventory levels within Snowflake and automatically issue purchase orders in an enterprise resource planning system when specific thresholds are met. This integration is supported by Snowflake’s open connectors initiative, which provides pre-built bridges to over 100 common enterprise software-as-a-service platforms.
In tandem with the intelligence upgrades, Snowflake Cortex Code has received significant technical enhancements designed to streamline the developer experience. Cortex Code now includes a multi-agent orchestration layer, allowing developers to build complex applications where multiple specialized AI models collaborate on a single objective. The update introduces improvements to Cortex Analyst and Cortex Search, which Snowflake reports have reduced inference latency by 35 percent compared to previous iterations. Furthermore, Cortex Code now supports a broader range of large language models, including the latest proprietary versions of Snowflake Arctic and optimized open-source models such as Llama 4 and Mistral Large 3.
Data governance remains a central pillar of the new releases. Snowflake integrated these AI capabilities directly with Snowflake Horizon, its built-in security and compliance solution. This ensures that all AI-driven actions and data access adhere to existing organizational policies, providing a lineage of thought log for every decision made by an autonomous agent. This transparency is intended to address enterprise concerns regarding the lack of visibility in AI decision-making processes.
Christian Kleinerman, Snowflake’s Executive Vice President of Product, highlighted that the new tools are designed to be accessible via low-code interfaces, enabling business analysts to build agents without extensive machine learning expertise. The company confirmed that these features are moving into general availability for customers in the Amazon Web Services and Microsoft Azure regions starting immediately, with Google Cloud support expected by the end of the second quarter.