Snowflake Plows Into GenAI With Slew of Announcements
- By Winston Thomas
- November 20, 2023
Snowflake, the Data Cloud company, is now extending its brand of data management simplicity to GenAI.
Its latest Snowday announcements seem to directly answer what data and AI teams have requested. It is also helping companies navigate the increasingly complex and seemingly fragmented world of GenAI development and LLMs.
One primary focus is governance, a key concern for everyone involved in AI or GenAI development. Snowflake is building on its previous efforts and governance framework to introduce the following products and features.
- Tip of the iceberg: Taking customer feedback on having no tradeoffs between managed and unmanaged Apache Iceberg tables, the company announced Unified Iceberg Tables (public preview soon). It gives users the choice of Snowflake-managed metadata or externally managed metadata. It unites all data in the Data Cloud.
- Reining in runaway budgets: With cost optimization in the minds of many data users, Snowflake's new Cost Management Interface (private preview) helps administrators stay in control. They can also set budgets and resource monitors to offer better oversight.
- Governance in the new Horizon: A significant concern for many companies is data and AI governance. Snowflake is rolling several announcements on the back of an expanded list of compliance certifications under Horizon, its governance solution. “Snowflake Horizon is Snowflake’s built-in governance solution with a unified set of compliance, security, privacy, interoperability, and access capabilities that’s provided to customers without additional configurations or protocols,” says Natalie Mead, vice president of APJ sales engineering at Snowflake at a press conference. Snowflake Horizon's Data Quality Monitoring feature (private preview) allows users to use provided metrics or create custom ones and email alerts to monitor data quality. A new Data Lineage UI (private preview) helps users trace how modifications made upstream can impact downstream. Other announcements include Differential Privacy Policies (in development) and Snowflake’s new UI-based classification workflow (public preview), Customer Classifiers (private preview) and international classification (generally available).
- Trust has a location on Horizon: The new Trust Center (private preview soon) allows users to monitor security in one place across clouds. While this offers better visibility and reduces account risk escalation, it can also reduce security monitoring costs.
Mead was careful to tout Snowflake’s approach to interoperability. "Snowflake Horizon makes it easy for customers to govern and take immediate action on data, apps, and more across clouds, teams, partners, and customers—both inside and outside of organizations. Integrations with other Apache Iceberg-compatible catalogs and engines and with partners further extend Snowflake Horizon across customers' ecosystems," she explains.
Leaping Into GenAI
GenAI is a big part of Snowflake’s latest push. But unlike other vendors simply bolting on GenAI capabilities, it is helping users to use GenAI to simplify LLM development.
The thread that connects the latest slew of GenAI-related announcements with its past efforts is Snowflake’s acquisition of Neeva. The latter’s GenAI search capabilities are now weaved into its products under a new fully-managed service called Snowflake Cortex (private preview). They include three unique components.
- Document AI (private preview): It brings Snowflake's 2022 acquisition of Applica to the forefront with the company's LLM. Users can discover and search documents easily using a visual interface and natural language.
- Universal Search (private preview): It uses LLM-powered search to hone in on the right data and apps. Users can search the Snowflake account, including databases, views, Iceberg Tables, data, and Snowflake Native Apps available on Snowflake Marketplace.
- Snowflake Copilot (private preview): It uses GenAI for Snowflake coding tasks. Users can now use natural language to query their data, write SQL queries on the fly, and do more advanced data analysis by asking the right questions or prompts.
Snowflake’s announcements set the foundations for a GenAI future. But it's interesting to see the slew of announcements highlighting how swiftly the company is moving to position itself in an increasingly crowded GenAI market.
“I attribute the swift progress in developing the five functions within Cortex to the exceptional team from Neeva. The recent advancements and impressive announcements at Snowflake events have largely been due to the team's influence after the acquisition,” says Mead.
These features and capabilities will matter as companies find their use cases and look for industry-based LLMs. Meanwhile, the company is identifying specific industry models and exploring its role in such developments while it continues integrating with open source LLMs.
Winston Thomas is the editor-in-chief of CDOTrends. He's a singularity believer, a blockchain enthusiast, and believes we already live in a metaverse. You can reach him at [email protected].
Image credit: iStockphoto/treasurephoto