Agentic AI Takes Center Stage at Snowflake Summit

Snowflake has been busy this year, steadily defining its position in the AI infrastructure race. At their annual Summit today, the company laid out an interconnected vision for the future of enterprise AI: a secure, agent-driven platform that simplifies how users interact with data.

Highlighting their build philosophy: “Building powerful AI apps and agents at scale hinges on enterprises having access to not only high-quality and relevant internal and external data, but also the business semantics of that data, in order to deliver trusted AI results,” said Prasanna Krishnan, Head of Apps & Collaboration and Horizon, Snowflake.

Clearly, this focus on AI Agents isn’t a sudden pivot. Last year Baris Gultekin, Snowflake’s Head of AI wrote: “2025 is when we will start seeing the hype of agentic systems start to bear fruit, with the first set of high-value agentic use cases going into production.”

The highlight of this push is a suite of agentic AI innovations, including Snowflake Intelligence and Data Science Agent, alongside enhancements to its Marketplace ecosystem. These are just the latest in a string of strategic moves to establish Snowflake as an Agentic AI enabler: moves that also include key acquisitions such as Crunchy Data.

Agentic Interfaces Replacing Traditional BI

Snowflake Intelligence, is a conversational interface that replaces traditional dashboards. Instead of building static reports, employees can now ask natural language questions and instantly receive answers, insights, or visualizations.

Built on top of Cortex Agents and large language models like Anthropic’s Claude, Snowflake Intelligence supports integration with enterprise tools such as Box, Workday, and Google Drive. It also taps into licensed third-party content through Cortex Knowledge Extensions.

“Snowflake Intelligence breaks down [these] barriers by democratizing the ability to extract meaningful intelligence from an organization’s entire enterprise data estate,” Gultekin said.

Cortex Knowledge Extensions make it possible to bring in real-time, licensed data from sources like USA TODAY, Stack Overflow, and The Associated Press: while preserving attribution and IP rights.

Data Science Agent for ML

Another new tool, Data Science Agent, aims to address one of the biggest bottlenecks in AI development: time-consuming manual ML workflows. This agentic assistant lives inside Snowflake Notebooks and helps automate routine machine learning tasks, from feature engineering to model training.

Data Science Agent breaks down ML development into verifiable steps, making it easier for teams to iterate and accelerate their time to production. It’s a clear response to the current state of enterprise AI, where teams spend too much time preparing data and too little time shipping models. With this tool, Snowflake is positioning itself as a productivity platform for data science.

Foundation: Cortex Agents

Underpinning both of these capabilities is Cortex Agents, introduced earlier this year and now in public preview. These agents form the foundational layer powering Snowflake’s agentic experiences.

Unlike traditional retrieval services, Cortex Agents can parse complex requests, determine next actions, and orchestrate data from both structured and unstructured sources like PDFs or spreadsheets. Powered by Claude 3.5 Sonnet, they already support real-world use cases such as fiscal calendar lookups, internal naming clarification, and priority indexing, without the need for complex configuration.

Expanding Into Industry-Specific Solutions

Snowflake’s agentic ambitions don’t stop at generic tools. In early May, the company unveiled its automotive data cloud initiative, a focused push into industry-specific AI applications.

With a 416% increase in manufacturing data apps since 2023, Snowflake is already supporting major players like CarMax, Nissan, Penske Logistics, and Subaru of New England. Use cases range from predictive maintenance to connected vehicle analytics.

As the auto industry undergoes transformation driven by electrification, autonomy, and Industry 4.0 practices, Snowflake offers a scalable foundation. With deep integrations with partners like Siemens and AWS and a growing marketplace of AI solutions, it’s betting that tailored data collaboration will drive real value for global manufacturers.

From AI-native assistants to compliance-grade data governance, Snowflake is laying the infrastructure for agentic systems that make it to production.