Announcement • Jun 04
Snowflake Announces New AI Innovations And Enhancements To Snowflake CoWork
Snowflake announced new AI innovations that expand how organizations build and operationalize AI across the business. The latest enhancements to Snowflake CoWork help enterprises move from insight to impactful action. Built for knowledge workers, CoWork is the personal agent that helps you work smarter. The latest CoWork enhancements are designed to put more AI capabilities into the hands of knowledge workers grappling with fragmented data, tools, and workflows that don’t understand the full context of their work. With new innovations including Artifacts, Cortex Sense, and personalization, CoWork unifies data, context, and AI into a single agent experience so that business teams can safely interact with AI to drive decisions and execution with confidence. At the same time, data teams can more easily unlock insights from data and train models at scale with the new Cortex Training. Snowflake brings AI directly to that data within a secure, governed environment. Snowflake gives customers the choice and flexibility to choose the AI models that best fit their business needs, with access to industry-leading models in Snowflake Cortex AI, Snowflake’s suite of AI products, from Anthropic, OpenAI, Google, Meta, Mistral, DeepSeek, and SpaceXAI’s models. CoWork builds a shared understanding of enterprise data that maps enterprise-wide business context with each user’s role to deliver tailored insights and actions. This includes turning everyday tasks into automated workflows with User Skills, and connecting directly to enterprise tools such as Google Drive, Salesforce, and Slack with Model Context Protocol (MCP) connectors, so teams can move from idea to action instantly across all of the systems they use. These MCP capabilities build on Snowflake’s intent to acquire Natoma, whose enterprise MCP platform will help provide secure connectivity, governance, identity-aware authorization, and auditability for AI agents operating across enterprise systems and tools. Deep Research comprehensively searches across all of a company's structured and unstructured data using multi-step reasoning and analysis. It operates within an organization’s governed data environment, exploring documents and business information to deliver fully cited, actionable insights that explain the critical “why” behind the numbers. Deep Research delivers performance across structured and unstructured enterprise data, outperforming single-agent systems by over a third on Snowflake's Hybrid Deep Research Benchmark. Snowflake is introducing Cortex Sense, a new capability that automatically brings together the data, business definitions, and operational knowledge that AI agents need to be trustworthy and useful. Cortex Sense works for both CoWork and Snowflake CoCo, the coding agent where you build faster, to unify that information into a shared context layer that agents can use immediately. Cortex Sense also comes with a series of prebuilt plugins tailored for various users’ roles and functions like finance and sales. These plugins combine skills, business logic, and MCP connectors to help enterprises quickly and easily deploy production-ready agents that are grounded in operational knowledge from day one. CoWork helps organizations move faster by using multi-agent orchestration to break down complex requests and coordinate specialized AI agents behind the scenes, without requiring users to select or manage them. Teams can automate their unique workflows with User Skills and leverage the new Skill Catalog to easily discover, share, and reuse them, enabling AI to more effectively support day-to-day tasks while scaling proven solutions and productivity gains across the organization. CoWork now continuously learns from individuals’ behavior with User Memory, allowing users to schedule tasks and get timely actions and recommendations tailored to their role through the tools they already rely on like Gmail or Slack. The new CoWork iOS mobile app allows knowledge workers to take action on their data from anywhere, as well as through Slack with CoWork Slackbot and a new Microsoft Excel Extension. Artifacts now include publishable dashboards with governed views of live, trusted data that knowledge workers can explore through conversation, not just clicks. Analysts can create and publish rich dashboards for knowledge workers to interact with in CoWork using natural language, reducing back-and-forth and accelerating decision-making. Teams can save, share, and collaborate on Artifacts, turning insights into reusable knowledge embedded in daily workflows, all with governance and access policies built in from the start. Snowflake is also introducing Cortex Training, extending Cortex AI to help organizations customize and train foundation models more easily with fully managed GPUs. Cortex Training enables enterprises to customize open-weight foundation models such as the Qwen or Mistral family of models to their own domain, data, and cost requirements without scrambling for scarce GPU capacity or worrying about high cost of GPU ownership due to low utilization. With Cortex Training, customers can train models where their data already lives, without managing distributed infrastructure or moving sensitive data to external environments, while completing up to two times more training runs for the same GPU budget based on initial customer assessments. Resolve AI made a multi-million-dollar commitment over two years to use Cortex Training to build domain-specific models through reinforcement learning (RL) on proprietary training data. Resolve AI uses Cortex Training to consolidate model training inside a secure, governed environment, accelerating work that previously required stitching together infrastructure across multiple platforms. Resolve AI offers domain-specific AI agents for production, built on multi-agent harnesses, simulated environments, and evaluation workflows. Cortex Training provides Resolve AI with governed infrastructure to run RL training workloads continuously, building and improving models that outperform general-purpose alternatives on the reasoning tasks that matter most to enterprise engineering teams in production. Snowflake engineering teams incorporate Resolve AI into their existing agentic workflows to run and manage production systems at scale.
