Snowflake’s AI Tool Promises to Make Your Data Talk—And It’s 90% Accurate

In June 2024, Snowflake launched Cortex when Christian Kleinerman, the Executive Vice President of Product, said, “If there are instances where we can deliver value to our customers sooner, faster, we’re not going to delay that.” Kleinerman’s statement highlighted Snowflake’s commitment to rapid value delivery. Today, the company has taken another significant step forward by introducing Cortex Analyst, an innovative agentic AI system designed to revolutionize self-service analytics.

Cortex Analyst, our fully managed service that provides a conversational interface to structured data in Snowflake, is now in public preview! Building reliable, self-service analytics applications that business users can trust has never been easier.

Trust is key for the adoption… pic.twitter.com/jvkmiZ9Nyu

— Snowflake (@SnowflakeDB) August 14, 2024

Leading provider of cloud data solutions, Snowflake, has revealed the release of Cortex Analyst’s public preview. Cortex Analyst is a fully managed service that gives companies a conversational interface for interacting with their data. With the help of this innovative technology, users may ask sophisticated business enquiries in simple English, and the AI system will take care of everything from translating prompts into SQL queries to producing precise responses.

Baris Gultekin, Snowflake’s head of AI, shared exclusive insights with VentureBeat, stating, “We’ve built systems that understand if the question is something that can be answered or ambiguous and cannot be answered with accessible data. If the question is ambiguous, we ask the user to restate and provide suggestions. Only after we know the question can be answered by the large language model, we pass it ahead to a series of LLMs, agentic models that generate SQL, reason about whether that SQL is correct, fix the incorrect SQL and then run that SQL to deliver the answer.”

The exceptional precision of Cortex Analyst is what makes it unique. According to Gultekin, the system provides insights with an accuracy rate of over 90%, which is a notable increase above current text-to-SQL solutions powered by LLM, including those from rivals like Databricks. This degree of precision is attained by combining several large language model (LLM) agents in an intelligent manner so that they operate together.

_{Source: Venture Beat}

The need for such a tool stems from the limitations of traditional business intelligence (BI) dashboards. While these dashboards have been the backbone of data analytics, they often prove rigid, making it challenging for users to drill down into specific metrics or obtain timely follow-up insights. Gultekin explains, “When you have a dashboard and you see something wrong, you immediately follow with three different questions to understand what’s happening. When you ask these questions, an analyst will come in, do the analysis and deliver the answer within a week or so. But, then, you may have more follow-up questions, which may keep the analytics loop open and slow down the decision-making process.”

Cortex Analyst addresses this issue by providing an intuitive, conversational interface that allows business users to instantly access the insights they need for critical decision-making. The system’s architecture is designed to ensure accuracy and eliminate hallucinations, grounding all responses in the enterprise’s data stored in the Snowflake data cloud.

To achieve this level of performance, Snowflake has leveraged a combination of its proprietary Arctic model along with models from Mistral and Meta. The system requires customers to provide semantic descriptions of their data assets during setup, enabling the LLM agents to understand the complete schema of a user’s data structure and provide context-aware responses.

“In real-world applications, you have tens of thousands of tables and hundreds of thousands of columns with strange names. For example, ‘Rev 1 and Rev 2’ could be iterations of what might mean revenue. Our customers can specify these metrics and their meaning in the semantic descriptions, enabling the system to use them when providing answers,” Gultekin added.

Currently, Cortex Analyst is available as a REST API, allowing developers to integrate it into various applications and customize how business users interact with the service. Snowflake also offers the option of using Streamlit to build dedicated apps with Cortex Analyst as the central engine.

During its private preview phase, Cortex Analyst was deployed by 40-50 enterprises, including pharmaceutical giant Bayer, to streamline their analytical workflows. The public preview is expected to significantly increase adoption, particularly as companies seek to leverage the power of LLMs for analytics without the associated implementation challenges and cost overheads.

Looking ahead, Snowflake has confirmed that Cortex Analyst will receive additional features, including support for multi-turn conversations to enhance interactivity and compatibility with more complex tables and schemas.

Sridhar Ramaswamy, CEO of Snowflake mentioned in his post, “Data isn’t just for data analysts and SQL experts. Too often, these teams bear the burden of being the go-to for non-technical users, creating a dependency that limits overall operational speed. To truly harness the power of data, it must be accessible to everyone.

Snowflake customers like Bayer are already seeing the benefits. By leveraging Cortex Analyst, they’re enabling self-serve analytics across teams, reducing reliance on IT, and accelerating data-driven decision-making.”