Latest Insights

Snowflake Summit Recap: The Year of Enterprise AI  

We are back in the real world this week after an productive week in San Francisco at Snowflake Summit, which means we get to take all the announcements and fanfare and get ready to apply them to the real-world opportunities and concerns of our data-minded clients.  

The Snowflake Keynote Stage was the epicenter of an onslaught of announcements, from innovation in data management with Dynamic Tables and Snowpipe Streaming to enhanced operability with Dynamic Tables and Iceberg Tables entering General Availability.  

And then there was AI – which should be no surprise to anyone who was at Snowflake Summit in 2023, or who hasn’t been living under a rock for the last few years. If 2023 was the Year of Talking About AI, 2024 is the Year of Enterprise AI. It was clear with every announcement that Snowflake has used Private and Public Preview releases to hone and mature its products and get them ready for prime time.  

Here are the announcements we are most excited about, and how we envision helping our customers deploy them immediately in their Snowflake environments. 

Snowflake Cortex AI 

There should only be two types of Snowflake users right now – those who are getting ready to use and/or build AI applications, and those who already are. Snowflake is working to democratize AI across its platform with releases that focus on creating experiences that allow organizations to interact directly with their data, while providing access to industry-best large language models (LLMs). Additional enhancements to Cortex AI include a fine-tuning feature that allows organizations to customize LLMs for a specific task using proprietary data, all within Snowflake.  

BOTTOM LINE: Get on the AI Train, today. Need help getting started? Contact us for a complimentary needs assessment and technology roadmap.  

Document AI 

Practically speaking, Document AI is something most organizations with troves of unstructured and semi-structured data should be able to use as soon as it is made available to the public (Document AI is currently available for private preview to commercial customers on AWS and Azure.) 

Document AI uses Arctic-TILT, an LLM developed by Snowflake, to extract data from PDFs, emails, reports and contracts – or any other text-intensive documents. With zero-shot extraction and fine-tuning, Document AI can find and extract specific information from previously unseen or unknown documents.  

BOTTOM LINE: Document AI will make an immediate impact for organizations that have had a hard time using unstructured data. If you have call-center transcripts or customer support documentation, you will soon be able to mine that data for better customer insights with Document AI. 

Relational AI  

While not a product release, were you even at the Keynotes if you didn’t hear about the technology darling of this year’s Snowflake Summit? It was exciting to learn more about Relational AI, and how enterprise organizations are bringing graph capabilities to the Data Cloud with the Knowledge Graph Coprocessor Snowflake Native App.  

BOTTOM LINE – For organizations just dipping their toes into AI, Relational AI’s knowledge graph functionalities can be a logical first step into graph analytics, rule-based reasoning and prescriptive and descriptive methodologies. Snowflake’s Marketplace Capacity Drawdown program makes this an enticing time to try something new.  

Iceberg Tables 

With the General Availability release of Iceberg Tables, Snowflake will make it easier for organizations to access data in any external cloud storage location for use in Snowflake. Iceberg Tables add flexibility to Snowflake deployment at scale, without incurring extra Snowflake storage costs.  

BOTTOM LINE: Iceberg Tables provide an elegant option for leveraging data that resides in data lakes outside of Snowflake.  

This is only a sample of the releases from Snowflake Summit 2024 – see a complete list of Snowflake product announcements here. If you’d like to learn more about maximizing your investment in the Snowflake Data Cloud, we can help.  

Return to Blog