03 April 2026 / 05:06 PM

March in the AI Data Cloud: Key Updates from Snowflake

Written by Michael Vollmer, Senior Executive and Snowflake Alliance Manager at SDG Group USA

Let’s be honest: the goal of every data leader right now is to reduce 'busy work' so their teams can focus on high-impact outcomes. This past month, Snowflake made that a lot easier. From autonomous agents to declarative infrastructure, the updates in March weren't just incremental - they were transformative. Here’s a quick look at the four big shifts I’m tracking for our customers and partners at SDG.

 

 1. The Dawn of Project SnowWork (PrPrv) 

Snowflake officially launched the research preview of Project SnowWork. This is a major pivot toward "Autonomous Enterprise AI." The goal is to move past simple chatbots and toward AI agents that can actually execute complex business workflows. It’s about outcome-driven AI that helps business users move faster without needing a PhD in prompt engineering.

  

Image courtesy of Snowflake

2. Cortex Code is Now Generally Available (GA)

For the builders out there, Cortex Code is now GA. This isn’t just another autocomplete tool; it’s now more integrated, supports native Windows environments via the CLI, and introduces Agent Teams. This allows you to break large coding assignments into parallel, coordinated tasks.

  

 Image courtesy of Snowflake 

3. Declarative Management (DCM Projects)

We’ve seen a huge shift toward "Data-Ops," and the preview of DCM Projects is a game-changer. It allows for a declarative, plan-then-deploy workflow for Snowflake objects. Essentially, you define the "end state" of your infrastructure in SQL, and Snowflake handles the "how." It’s a massive win for version control and multi-environment stability.

High-level workflow for managing a DCM project:

  

 

 Image courtesy of Snowflake 

4. Scaling Cortex Search

Cortex Search saw big upgrades this month, including multi-index search (targeting specific columns in one service) and custom vector embeddings. For our clients looking to build RAG (Retrieval-Augmented Generation) applications, this provides the granular control needed for enterprise-grade accuracy.

  Image courtesy of Snowflake

 

Learn more about our partnership with Snowflake and how our team can help you establish a strong foundation for innovation in the cloud.

Related Insights & News