Snowflake debuts Cortex Analyst, a fully managed service for self-service analytics

Snowflake is launching Cortex Analyst, a new fully managed service that enables users to interact with their structured data using natural language through a conversational platform, in public preview. Cortex Analyst aims to democratize data analytics by giving developers the tools to build conversational applications that can deliver near-real-time answers to natural language queries quickly, thus lifting the pressure off data analysts who are often in charge of interpreting the BI dashboards and reports containing the answers to their organizations' data questions.

To boost accuracy, the service is built on a combination of models from Mistral and Llama. Benchmarking reveals that Cortex Analyst is about twice as accurate as state-of-the-art models and 14% more accurate than a competing data analysis solution. Moreover, to ensure customers Cortex Analyst was built to guarantee the privacy and safety of customer data, Snowflake states that it does not train its system on customer data, that the service integrates Snowflake's privacy and governance features, and that customer data never leaves Snowflake's governance boundary. Finally, the company has also ensured that Cortex Analyst allows for flexibility; since it is available as a REST API, the service can be integrated into any application.

The service is already making waves, with pharmaceutical giant Bayer using it to power self-serve analytics on their enterprise data. As AI fuels an ongoing transformation of data accessibility, Cortex Analyst represents a significant step towards more intuitive and accurate data interactions for non-technical users. Snowflake plans to expand Cortex Analyst's capabilities in the coming months, including support for multi-turn conversations and integration with Cortex Search.