Nabugu - stock.adobe.com

AI tools highlight latest swath of Snowflake capabilities

The vendor is developing a natural language interface that unites disparate data types and an agent that makes data scientists more efficient, among other features.

Snowflake on Tuesday introduced a broad array of capabilities spanning each of the vendor's areas of focus, including AI tools that simplify development and analysis, and data engineering features aimed at making structured and unstructured data available together.

Snowflake unveiled the new capabilities during Summit, its user conference in San Francisco.

In addition to AI and data engineering tools, Snowflake revealed platform capabilities to optimize performance, analytics capabilities that simplify queries, and features aimed at simplifying and improving collaboration.

Collectively, the new features -- most of which are in preview -- are in line with market trends and address the needs of users, according to David Menninger, an analyst at ISG Software Research.

The announcements cover a broad range of topics that align with developments in the market. They indicate significant and ongoing investment, which should make Snowflake users happy.
David MenningerAnalyst, ISG Software Research

"The announcements cover a broad range of topics that align with developments in the market," he said. "They indicate significant and ongoing investment, which should make Snowflake users happy."

Based in Bozeman, Mont., but with no central headquarters, Snowflake is a data cloud vendor that has expanded into AI development since OpenAI's November 2022 launch of ChatGPT marked a significant improvement in generative AI (GenAI) technology and sparked a surge in AI development.

New AI capabilities

While Snowflake continues to add data management capabilities, such as the recent update of its AI Manufacturing Data Cloud, much of the vendor's product development since Sridhar Ramaswamy was named CEO in February 2024 has centered on AI.

Most significantly, the vendor developed Cortex AI, an environment within Snowflake that provides customers with AI-powered capabilities and the tools to build and deploy their own AI and machine learning applications.

The vendor's latest AI innovations include Snowflake Intelligence and Data Science Agent.

Snowflake Intelligence, not yet in public preview, is a conversational AI interface designed to change the way users interact with data.

Built with large language models (LLMs) from Anthropic and OpenAI and powered by agentic AI capabilities from Snowflake, Snowflake Intelligence brings structured and unstructured data together to enable users to query and analyze the full breadth of their organization's data without writing code. In addition, because it runs within customers' Snowflake environment, security, data masking and governance policies are automatically applied.

Data management and analytics users have long wanted natural language interfaces, according to Menninger. GenAI finally made them possible. As a result, Snowflake Intelligence is a significant addition.

"Line-of-business personnel have been clamoring for natural language processing for years," Menninger said. "Nearly every analytics vendor has added copilots or assistants, and data platform vendors are now adding them as well."

Data Science Agent, meanwhile, is an agentic AI assistant not yet in private preview that aims to make data scientists more efficient by enabling them to automate routine model development tasks using natural language.

The agent uses Anthropic's Claude LLM to break down machine learning workflows into steps, such as data preparation and feature engineering, and applies techniques such as multistep reasoning and contextual understanding to carry out the steps.

Both the new AI capabilities are valuable for Snowflake users, according to Kevin Petrie, an analyst at BARC U.S.

"These agents offer significant new capabilities for Snowflake users," he said.

Beyond their value for Snowflake users, the new AI capabilities demonstrate a meaningful approach to AI that differs from rival Databricks, according to Petrie. While Databricks is focused on enabling highly skilled data scientists to build custom agents for complex AI and machine learning development, Snowflake is focused on helping analysts and data scientists use premade AI tools and templates to analyze diverse data sets.

"This announcement reinforces Snowflake's differentiated strategy and distinct target market vs. the competition," Petrie said.

Additional new capabilities

While Snowflake Intelligence aims to speed and simplify analysis, new data engineering capabilities that enable access to unstructured data in addition to structured data help make it possible.

Snowflake Openflow, now generally available on AWS, is a multimodal data ingestion service that enables users to connect to nearly any data source and then integrate their data within Snowflake. By automatically ingesting and unifying different types of data and formats, Openflow is designed to save hours of manual labor.

Perhaps most significant is that Openflow enables users to combine the structured data that has historically informed analytics tools with the unstructured data that now makes up the vast majority of all data, according to Christian Kleinerman, Snowflake's executive vice president of product.

"The most important thing for me to highlight is that Openflow is making both structured data as well as unstructured data available to Snowflake," he said during a virtual media conference May 28.

Other new Snowflake capabilities include the following:

  • The general availability of Generation 2 of its Standard Warehouse, which includes improved query speeds.
  • Adaptive Compute, a feature in private preview that intelligently selects the right amount of compute power required to carry out Standard Warehouse workloads to optimize performance and cost.
  • SnowConvert AI, an agentic AI-powered tool that simplifies data migration from other data management platforms to Snowflake.
  • Cortex AISQL, a feature in public preview that will automatically be applied to data as it's loaded into Snowflake, so users can run AI-powered SQL queries on all data types without having to manually restructure data.
  • Cortex Knowledge Extensions, a tool soon to be generally available on the Snowflake Marketplace that lets users enrich AI applications and agents with unstructured data from third-party providers.
  • Sharing of semantic models within organizations through Snowflake's Internal Marketplace or with third parties through the Snowflake Marketplace.
  • Agentic Snowflake Native Apps on Snowflake Marketplace so users can share and monetize the agents they build with Snowflake.

Petrie highlighted the importance of improved query speed and Adaptive Compute.

"The compute innovations address the big three requirements of reliability, performance and ease of use," he said, noting that BARC research shows that AI adopters prioritize those requirements when evaluating data tools and platforms.

Menninger, meanwhile, noted the importance of sharing semantic models to make them more interoperable.

"Sharing data is great, but if you don't share the metadata, it really makes it more difficult to use the data effectively," he said. "As an industry, we've gotten much better at sharing data. Now we need to tackle the problem of sharing metadata."

Looking ahead

Given that most of the new features Snowflake introduced are in the preview or testing phases, they essentially represent the vendor's near-term product development roadmap.

However, as Snowflake evolves, agentic AI will play a greater role in its product development plans, according to Ramaswamy. In addition, the underlying focus guiding the vendor's plans is to provide the tools to make customers' data ready for AI.

"AI represents a new way of thinking about how different functions within a company operate," Ramaswamy said during the press conference. "The role Snowflake plays is to be an ally, an agent of transformation for all of our customers."

As it aids customers' transformation from traditional analytics to AI-driven analysis, Snowflake's focus needs to be on making some key features that are in preview generally available, according to Menninger. In particular, Cortex Agents, a managed service designed to simplify agentic AI development, could help Snowflake's customers.

"Agentic AI is the hot topic right now, but enterprises are hamstrung because many of the agentic features in vendors' platforms are not yet fully supported," Menninger said.

Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.

Dig Deeper on Data integration