Snowflake has rooted its offerings more deeply in artificial intelligence, betting that its data cloud platform can deliver the promise of the technology at a time when many organisations are reappraising their approach to AI implementation.
Among a flurry of new service announcements made at the end of last year, Snowflake unveiled plans to launch an agentic AI solution called Snowflake Intelligence. Currently in preview, the product will enable clients to better interrogate and use their own enterprise data by utilising tools and capabilities on the platform, the company said.
Along with plans to enable clients to build internal data products marketplaces, and after forging a new collaboration and another acquisition, the company says its latest announcements will provide clients with a full-suite of cloud-based tools to tackle new and emerging use cases.
“Our customers face the dual challenge of driving innovation while staying compliant; they want best-in-class tools without uncompromising governance and security,” Rinesh Patel, Snowflake global head of industry, financial services, told Data Management Insight.
Future Proofing
Snowflake is investing heavily into AI even amid a waning of the initial excitement about the potential of AI that followed the launch of generative AI-based ChatGPT in late 2023.
Experts across the data spectrum have warned that many organisations are slowing their initial rush to implement AI as the cost-benefits of their investments appear less rosy than imagined and integration proves more difficult than hoped. Snowflake’s principal data strategist Jennifer Belissent added to the sense of caution, telling media recently that she thought 2025 would be the year in which GenAI must demonstrate its true value.
Snowflake believes it has the capabilities to bring AI’s potential to enterprises. The company’s confidence was demonstrated this week when it unveiled a US$20 million investment into an upskilling programme called One Million Minds + One Platform. The scheme aims to train 100,000 people and upskill a million more by 2027 in the capabilities needed to drive AI.
Commercially, the biggest recent demonstration of its commitment to AI came with the announcement of Snowflake Intelligence, which the company says will streamline workflows through the use of AI agents dedicated to automating specific tasks.
Clients will be given the tools to build individual agents tailored to their needs that can run within their Snowflake accounts, leveraging third-party data tools and Snowflake products to help them access, interrogate and optimise the use of their enterprise data.
“Through Snowflake Intelligence we’re enabling the ability to deploy agents that distinguish between different data tasks, for example, chatting with unstructured or structured data, and then autonomously deploy the right tool for the task,” said Patel. “This is the next frontier in democratising access to data and AI tools for business users.”
Partnership and Innovation
The company’s fully managed GenAI-based Cortex AI is the motor behind Snowflake Intelligence, enabling clients to run queries as well as access and use analytical models more easily. Through integration with existing capabilities on the platform, Snowflake said that clients will also enjoy “strong enterprise-grade compliance, security, privacy, discovery, and collaboration capabilities”.
Snowflake Intelligence will be complemented by a strategic partnership with AI safety and research company Anthropic, which will enable the deployment of the latest Claude 3.5 Sonnet model into the Snowflake platform to “unlock state-of-the-art LLM powers for our enterprise customers,” said Patel.
The new solution will also benefit from Snowflake’s acquisition of data pipeline accelerator and observability provider Datavolo. This, said Patel, would help clients integrate unstructured data.
Internal Marketplace
Another innovation under preview is an internal marketplace that will enable clients to make their own data, models and other data products available across their enterprise.
“Many customers have been forced into a ‘build-first’ approach to self-service, wrestling with fragmented data lakes, data mesh architectures, multiple catalogues and lineage tools,” said Patel. “It’s been a slow, painful process—delivering far less value than expected and delaying the business outcomes they urgently need.”
The internal marketplace will let customers centralise and share their directory of data, applications, AI, and machine learning models, he said. Teams can seamlessly browse, discover, and access these assets across the organisation, “unlocking collaboration at scale”.
“You’re going to have all these different teams, which have their own versions of their data and data models, and now they’re all going to converge on an internal marketplace where they can find things, try things and share things in a more collaborative way.”
Subscribe to our newsletter