Snowflake used its annual Summit conference to unveil a new set of AI features aimed at making enterprise adoption easier while keeping data controls in place.
At the company’s keynote in San Francisco, Snowflake outlined updates across its AI platform that focus on governance, context, automation and fresh data access. The announcements are designed to help businesses use AI agents and coding tools with more of their own operational knowledge, while limiting security risks.
One of the main updates is Snowflake CoWork, the renamed version of Snowflake Intelligence. The company is positioning it as a personal AI agent for knowledge workers that brings together data, context and tools in one system.
Snowflake said CoWork now includes Cortex Sense, which combines data with business definitions and operational knowledge. It also adds User Memory, a feature that lets users schedule tasks and take actions with agents through services such as Gmail and Slack. Another addition, Skill Catalog, is meant to help employees discover, share and reuse agent capabilities.
Snowflake also introduced Horizon Catalog as a central hub for AI governance, context and security. The company said Horizon Context is intended to ensure that people, tools and AI agents rely on the same business definitions. Adaptive Compute is meant to automatically manage compute resources, while new security features include verified agent identities, continuous security posture management and defenses against jailbreak attempts and zero-day vulnerabilities.
Snowflake also renamed Cortex Code to Snowflake CoCo. The coding assistant can now carry out tasks on its own without requiring the user to remain active on screen, according to the company.
Another product announcement was Datastream, which Snowflake says removes the need for data brokers, connectors or streaming infrastructure. The company says that will make it easier for AI applications to keep access to current data.
Christian Kleinerman, Snowflake’s executive vice president of product, said the company’s goal is to help organizations become more productive while also giving them confidence around security, compliance and governance.
Snowflake also highlighted progress in its partnership with Anthropic, including growing use of Anthropic’s Claude models within Snowflake Cortex AI, the company’s AI product suite.
In the opening keynote, Anthropic President Daniela Amodei said the company’s focus on enterprise customers is tied to trust, describing trust as something that helps businesses move faster. Her comments came during a conversation with Snowflake CEO Sridhar Ramaswamy.
Snowflake’s latest releases appear aimed at two major challenges facing enterprise AI adoption: making tools easier to use and making them safer to deploy. By reducing the need for extra infrastructure and adding more guardrails around access and governance, the company is positioning itself as a platform for businesses that want to use frontier AI models with tighter control over their data.
The announcement comes as companies across industries continue to look for practical ways to integrate AI into daily workflows without exposing sensitive information or losing oversight of automated systems. Snowflake’s pitch is that its platform can help enterprises do both.