Snowflake has unveiled Project SnowWork, an autonomous AI interface designed to carry out enterprise tasks without requiring data team intervention. The level of automation is ambitieus, with everything from forecasts to reports being AI-generated. Built on its AI Data Cloud, Snowflake Intelligence, and Cortex Code, the offering aims to bring Snowflake directly to business users rather than staying behind the scenes.
The project is seemingly in quite early stages still. Snowflake emphasizes the tooling is currently under development. The company describes SnowWork as an autonomous layer for its data cloud. Rather than simply answering questions about data, SnowWork is designed to do the actual work, producing forecasts, churn analyses, supply chain reports, and even presentation slides on behalf of business users in finance, marketing, and sales.
The project combines several existing Snowflake technologies. These consist of the AI Data Cloud, Snowflake Intelligence, and Cortex Code. Bala Kasiviswanathan, VP of developer and AI experiences at Snowflake, describes the thinking behind it plainly: “The idea is to have AI act more like a proactive collaborator. So instead of just asking questions, business users across functions like finance, marketing, or sales can ask for outcomes.”
it’s a more ambitious set of capabilities than Snowflake has shown before. Last November, it made Snowflake Intelligence generally available. And in February, the company signed a $200 million partnership with OpenAI to deepen its AI Data Cloud capabilities, relying on third-party AI power to do so. SnowWork appears to be the next logical step in that escalating investment.
From back-end utility to front-office tool
The strategic angle is fairly straightforward. Snowflake wants to move from being infrastructure that most employees never touch to a productivity layer that hundreds use daily. Today, a typical business user doesn’t log into Snowflake directly. Instead, their experience comes filtered through a BI tool. SnowWork would change that dynamic entirely.
That ambition also puts Snowflake in more direct competition with Microsoft, Google, and Salesforce, all of which are building AI layers aimed at the same front-office territory. Snowflake CEO Sridhar Ramaswamy has previously warned that software risks becoming a “dumb data pipe” for AI labs. SnowWork is in part a response to that concern.
Kasiviswanathan pitched SnowWork as more accurate than generic chatbots because it runs on secured, governed enterprise data rather than relying on manual coordination.
Open questions remain
Still, SnowWork is not yet available and carries several unanswered questions. No launch date or pricing model has been announced. If the product compresses analysis from weeks to minutes but significantly increases Snowflake consumption costs, finance leaders will have to do the math carefully. And as Snowflake’s own AI progress shows, moving from assisted analysis to genuinely autonomous output is a non-trivial step that requires consistent accuracy before enterprises will fully trust the results.
Snowflake is currently testing SnowWork with select customers and has not announced a broader rollout timeline.