Empowering Data Insights: Snowflake Summit Takeaways


Now that our Unifiers have had a few weeks to enjoy the warm season, we thought it was time to get our minds a little chillier and reflect on the incredible week at Snowflake Summit 2023. We are excited to share some key insights about the advances that are shaping the future of data analytics.

Here are our top takeaways:

01 / Clean-up to be ready for AI. Data quality and a well-defined data strategy are crucial for effectively leveraging AI technologies like Large Language Models (LLMs). Your strategy must involve ensuring clean data and actively managing your contextual metadata to keep it up to date.

02 / Iceberg tables are cool. Iceberg tables are a new open-source Apache format for external tables that provide better performance compared to current Snowflake external tables. With Iceberg tables, you can query data against your data lake using relational databases without having to move the data into Snowflake, thus addressing concerns regarding vendor lock-in.

03 / New applications to look out for in the Marketplace. One of the most exciting offerings is an “auto-ETL” for the application’s data, which can help reduce development time and overcome initial data catalog challenges. ServiceNow is the first approved application in the Marketplace. You can sign in with your ServiceNow credentials, and the app will build the appropriate objects in your Account and transfer the data to you. Other applications, such as Hex, can alter the interface and provide a collaborative Python notebook environment for data exploration and analysis.

04 / In-the-box processes spur out-of-the-box performance. Containerization processes enable running containers on Snowflake, providing flexibility and optimization. Your container processing can utilize standard Snowflake compute clusters or be switched to NVIDIA GPUs if your application is optimized for GPUs. This is particularly beneficial for AI and machine learning workloads.

05 / Training your AI models. No leash required. Document AI focuses on processing unstructured, digitized documents, such as contracts, legal forms, healthcare forms, or any other type of unstructured text. It utilizes a universal zero-shot model, meaning it does not need to see your data first, to train AI models using natural language processing. You only need to ask questions on a subset of your documents to train the model, and then the model can be applied to all documents to create structured data tables.

06 / Get more dynamic. Dynamic tables are no longer in preview and are available for use, offering more flexibility in data manipulation and management.

07 / Cash in on snow. You can use your Snowflake credits in Snowflake’s Marketplace to pay for applications, enhancing the flexibility of resource allocation.

08 / More access. Less burden. External network access from Snowpark allows teams to directly utilize APIs from Snowflake, expanding possibilities without relying on external teams.

09/ Interfaces get a glow-up. Streamlit native integration enables the creation of visually appealing interfaces and dashboards within the Snowflake environment.

Want to dig in and learn more about the latest Snowflake innovations? Recorded sessions from the Snowflake Summit are now available On-Demand.

Want to understand how Unify and Snowflake can help bring the benefits of migrating data to the cloud to life, including faster speed-to-insight and reduced cost? Learn more about our partnership.