AWS re:Invent 2025 - Explore what’s new in data and AI governance with SageMaker Catalog (ANT308)

Summary of AWS re:Invent 2025 - Explore what's new in data and AI governance with SageMaker Catalog (ANT308)

Key Highlights

• Importance of data and metadata management for successful AI initiatives
• Introduction to Amazon SageMaker Unified Studio and SageMaker Catalog
• New features and capabilities to enable metadata-driven data discovery and governance

Data and Metadata as the Foundation for AI

• Data is the foundation for AI, but managing data correctly is crucial
• Key challenges identified by customers:
  • Need for a single place to catalog all data, models, dashboards, and assets
  • Requirement for rich metadata to provide context and enable discovery
  • Consistent governance and access control across all data assets

Amazon SageMaker Unified Studio and SageMaker Catalog

• SageMaker Unified Studio provides a centralized platform to catalog and manage all data assets
• SageMaker Catalog is the metadata-driven heart of the solution, enabling:
  • Unified view of structured and unstructured data, models, dashboards, and more
  • Automatic metadata generation and association of glossary terms
  • Enforcement of metadata rules and policies

New Capabilities in SageMaker Catalog

1. Column-level Metadata: Ability to add context and governance at the column level, not just table or asset level
2. Automatic Glossary Term Suggestions: SageMaker Catalog can automatically suggest and apply relevant glossary terms to new data assets
3. Metadata Sync with External Catalogs: Ability to sync metadata from non-AWS catalogs (e.g., Databricks, Snowflake) into SageMaker Catalog
4. Catalog Federation with Iceberg Catalogs: Capability to federate metadata from external Iceberg-compatible catalogs into SageMaker Catalog
5. One-click Onboarding of Existing Data Assets: Seamless integration of existing data assets into SageMaker Catalog, leveraging existing IAM permissions

Demonstration and Customer Example

• Demonstration showcasing the end-to-end workflow:
  • Data engineer creating a data pipeline using SageMaker Studio
  • Data steward documenting the data asset with business context and metadata
  • Data analyst exploring the cataloged data and creating a dashboard
  • Data scientist leveraging the new SageMaker notebook experience to build a forecasting model
• Customer example from Nat West Bank:
  • Nat West's journey to modernize their data architecture and enable AI-powered banking
  • Challenges of managing data in a 300-year-old organization and the importance of metadata and governance
  • How Nat West is leveraging SageMaker Unified Studio and Catalog to federate data access and drive reuse

Key Takeaways

1. Extend your data architecture to incorporate metadata and governance as a foundational layer to enable AI initiatives.
2. Leverage SageMaker Unified Studio as a central platform to enable collaboration and innovation across different personas (data engineers, data stewards, data analysts, data scientists).
3. Start small, build incrementally, and deliver ongoing value to gain organizational support for your data transformation journey.