Snowflake launches desktop AI agent for enterprise workflows

Christian Kleinerman, executive vice president of product at Snowflake, speaks during the company’s media event in Seoul, Thursday. Courtesy of Snowflake

U.S.-based software company Snowflake launched Project SnowWork, an artificial intelligence (AI) assistant designed to automate enterprise workflows, in a research preview on Thursday, marking the latest step in its push to bring agentic enterprise AI directly to business users’ desktops.

“The key insight (of the platform) is we will be introducing ‘profiles’ … There is a profile for product managers, sales leaders and finance. And it personalizes every enterprise context to that specific role with the right data, right skills and right insights,” Christian Kleinerman, the company’s executive vice president of product, said during a media event in Seoul's Jung District.

“We are incredibly excited to work with customers to shape what we think is the future of work and the future of the agentic enterprise.”

Project SnowWork allows business users to request tasks in natural language and have AI autonomously plan and execute multi-step workflows — from data analysis to report writing and presentation generation.

By combining multi-step task execution, persona-specific skills and built-in security and access controls, it is designed to act as a trusted agentic partner that can safely act on enterprise data rather than just summarize it.

The launch reflects Snowflake’s strategy for a broader enterprise AI stack, which already includes Snowflake Intelligence agent for natural-language analytics and Cortex services for building and deploying customized AI tools.

The software company provides a single environment where enterprises can store, process, manage, analyze and share data, and even build applications and AI workloads on top of this data foundation. It currently serves more than 13,300 customers globally.

With this new addition, users can move from intent to execution without filing tickets with data teams or waiting on static dashboards, while developers can use tools, such as Cortex Code, to productionize new AI workflows inside the same governed Snowflake environment.

“Our goal is to help companies of all sizes, all industries, to make AI real and productive for all users,” Kleinerman said.

“It is for everyone in the company to leverage all the business logic they have and help them think through structured data, semi-structured data, unstructured data and streaming data.”

Since launching in 2021, Snowflake Korea has rapidly expanded its footprint, signing up around 80 percent of Korea’s top 10 conglomerate groups and growing local platform consumption more than ninefold over the past four years.

The company highlighted e-commerce operator Lotte ON, which used Snowflake’s AI data platform to break down data silos and power AI-driven use cases such as customer segmentation and real-time product recommendations, delivering a 32-percent reduction in operating costs and a 40-percent improvement in overall performance.