Snowflake CIO identifies AI focus in 2023 data trends report

Snowflake got its start by bringing data warehouse technology to the cloud, but now in 2023, like every other vendor, it finds artificial intelligence (AI) permeating nearly every discussion.

In an exclusive interview with VentureBeat, Sunny Bedi, CIO and CDO at Snowflake, detailed the latest findings from his company’s 2023 Data Trends Report, which is being released today. The report reveals that (surprise, surprise) AI is top-of-mind and a foundational use case for a growing number of organizations that use Snowflake.

In his role as CIO, Bedi has a front row seat into not only how other enterprises use Snowflake, but how Snowflake itself uses data and AI to advance its business.

Overall, Snowflake now sees four key trends:

1. Companies are connecting data everywhere they can.
2. State-of-the-art companies are bringing their work to the data — not vice-versa.
3. Governance matters more now than ever before.
4. Companies are increasingly embracing automation.

“Year over year, we have seen a 207% growth of data coming into Snowflake across the three cloud providers, AWS, GCP and Azure,” Bedi told VentureBeat. “With that we’re seeing more computational workloads that need building with advanced tools, and if organizations don’t connect to a single data source of truth, they will fall behind.”

Bringing code to data is critical

However, simply having a single source of truth for data isn’t enough for organizations to actually be able to benefit from data. That’s where a new era of programmability and AI comes into play.

In 2022, Snowflake announced its Snowpark framework for data science and application development, which is all about bringing code to data. The primary language for code is Python; Bedi noted that 88% of the jobs that run on Snowpark are written in Python rather than any other language such as SQL or Java.

“We see an increased adoption of how Snowpark is allowing code to move to the data rather than the other way around, and as such, speed and governance [are] becoming incredibly efficient,” Bedi said.

There is also an intersection between Snowpark and Streamlit. Snowflake acquired the Streamlit technology in March 2022 to help with application development. And overlaying the development capabilities is the growing world of generative AI and its potential to bring a new interface to code and data.

Snowpark, ChatGPT and the end of business dashboards

Snowflake itself is using Snowpark and AI to help improve its own operations. 

Bedi recounted that Snowflake’s CEO Frank Slootman asked him to build an application that would make Slootman’s life a bit easier. What the CEO wanted was to be able to type a simple natural language question in English about some aspect of Snowflake’s business and operations, and get a response.

Snowflake, like many other businesses, relies on business intelligence dashboards to help provide management with key performance indicators and metrics.

“With ChatGPT, Snowflake and Streamlit, we built an incredibly easy application for him in two days, where he can go and ask questions about sales and other types of metrics that he is interested in,” Bedi said.

Snowflake is not yet commercially offering this integration with ChatGPT/generative AI to its users — yet. Bedi emphasized that the small application his team built was for an internal use case. That said, he hinted that it could be productized in the future, and the larger trend is that the programmability of data inside of Snowflake now makes new AI-powered use cases possible.

Just last week, Snowflake announced its acquisition of AI search vendor Neeva, which will be a major component of the company’s future AI services.

“You’re going to have an ability to use next-generation search technology powered by large language models using Neeva,” Bedi said. “This will also enable Snowflake users and app developers to build rich, search-enabled conversational experiences.”