Snowflake Inc.’s raison d’etre is eliminating the need for data silos through a unified platform that simplifies data architectures and enables congruent governance controls.

But it’s not enough to consolidate, secure and perform normal analysis on the data anymore. With artificial intelligence expanding the playing field for data programmability, what are solutions providers doing to bring AI into that data platform without any governance and compatibility tradeoffs?

“…The focus originally was let’s help our customers break down silos,” said Christian Kleinerman (pictured), senior vice president of product at Snowflake. “The dream of having a single data fabric where you can ask questions about your customer holistically, your product line holistically. And what we saw was, okay, on one side you’re un-siloing and on the other side you are re-siloing. One of the biggest reasons for re-siloing is applications because [with] every application [you] realized with more data, you do better. Our goal is to keep customers un-siloed — bring applications, bring business logic to the data.”

Kleinerman spoke with theCUBE industry analysts Dave Vellante and  George Gilbert at Snowflake Summit, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed Snowflake adapting to modern AI-driven enterprise data applications. (* Disclosure below.)

New managed/unmanaged tables and GenAI for governance

Cloud is evolving to meet users where they are. This hybrid approach is changing how companies construct their cloud infrastructure across on-premise, public/private cloud and the edge.

The Iceberg idea is a frictionless deployment where customers aren’t forced to abandon expensive existing installations and, instead, simply adapt what’s already there, according to Kleinerman.

“Iceberg is our way to meet customers where they are,” he said. “If someone already invested in having tens or hundreds of petabytes of data in parquet files in cloud storage, I don’t know that we need to force anyone down any one path. One of the big announcements today is these two modes of iceberg unified tables. If you just want to start querying data that you have in cloud storage, unmanaged tables are good for you — It’s a read-only activity. But if you want to say Snowflake should take over the administration, management and consistency of that data, you can upgrade it to a managed table.”

On the ingest side of things, Document AI is Snowflake’s new push in the realm of generative AI. It allows enterprises to swiftly glean value from their swathe of heterogeneous documents and files.

Underpinned by Applica’s breakthrough technology, Document AI provides an easy way for enterprises to mobilize crucial unstructured information that is oftentimes spread out across disparate silos, according to Kleinerman.

“The purity of our design makes things composable and by virtue of that, usable in a variety of contexts,” he explained. “We show the user interface based off Document AI but there’s also a function that you can call. That function can be included in a dynamic table or in a pipeline that you build yourself.”

Generative AI, from a Snowflake viewpoint, is another way to transform data into value. The company’s chief differentiator in the space is in applying GenAI to data governance, Kleinerman added.

“I would say the broad approach that we’ve been pursuing around bringing computation to the data is all about, ‘let’s do rich programmability of data without compromising governance.’ — because we provide lots of capabilities to deliver governance.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of Snowflake Summit:

(* Disclosure: TheCUBE is a paid media partner for Snowflake Summit. Neither Snowflake Inc., the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE