With the new features Snowflake Intelligence and Data Science Agent, Snowflake aims to make AI more accessible for business users and data scientists.
During its annual summit in San Francisco, Snowflake announced two new AI features: Snowflake Intelligence and Data Science Agent. Both solutions are designed to make it easier for business users and data scientists to deploy AI and machine learning (ML) within their existing data platform.
Snowflake doesn’t want to simply jump on the AI agent hype bandwagon: the company prefers to call the new solutions “data agents”. The announced features will first appear in public or private preview.
Snowflake Intelligence: Data in Human Language
Snowflake Intelligence offers business users and data analysts the ability to gain insights from both structured and unstructured data using natural language. Intelligence is an AI layer that Snowflake applies over its data cloud. The feature works within the existing Snowflake environment and supports integrations with external platforms such as Google Drive and Workday. Thanks to this feature, users can generate visualizations and take actions without writing code.
read also
Snowflake Intelligence Converts Data into Human Language
Under the hood, Snowflake Intelligence works with Cortex Agents, powered by models from OpenAI and Anthropic, among others. The tool maintains existing security and governance settings within Snowflake so that users can only see data they are authorized to access. Cortex Extensions, available through Snowflake’s AI marketplace, expand search capabilities to external knowledge sources.
The chat capability Snowflake introduces closely resembles the new functionalities in Microsoft Fabric. As Microsoft increasingly involves itself in AI and data analytics, acting as a competitor alongside being a partner, Snowflake is stepping up its game.
ML for Data Scientists
Snowflake doesn’t lose sight of the data scientist. Data Science Agent automates various parts of the ML development process, such as data analysis, feature engineering, and model training. The goal is to reduce time spent on repetitive tasks so that data scientists can bring models into production faster. The agent generates working ML pipelines based on natural language commands and supports iterative improvements.
Both features are designed to help companies make decisions faster based on their data, without technical barriers. This should narrow the gap between data analysts, AI developers, and business users.
AI-Ready
The AI fun doesn’t stop there during Snowflake Summit. Snowflake introduces the ability to share semantic models. This allows companies to enrich AI applications with data from internal teams or external providers that are “AI-ready”. This way, companies don’t have to develop models themselves, which increases the reliability and consistency of AI answers.
Finally, Snowflake also launches Agentic Native Apps via the marketplace. These are applications built on agentic AI that companies can use directly within their data environment. Vendors can share these apps, while consumers can easily discover and deploy them without moving their data.