AWS re:Invent 2025 - Strategy First: Empowering the Business to Lead in the AI Era (AIM348)

Empowering Business Teams to Lead in the AI Era

Enabling Business-Driven AI Innovation

• The presentation emphasizes the importance of empowering business teams to drive AI solutions, rather than relying solely on IT or data teams.
• The speaker, Adrian Estala (COO at Starburst), and Raj Palen Swami (from MUFG) discuss a framework for bridging the gap between business teams and AI capabilities.

Three-Layer Foundation for Business Enablement

1. Data Access:

   • Allowing business teams to access data where it resides, whether in the cloud or on-premises, without the need for data migration.
   • Leveraging data federation capabilities to connect to various data sources.

2. Semantic Layer:

   • Creating a business-facing semantic layer that organizes data into logical, easily understandable data products.
   • Aligning these data products with the business's core products and functions, enabling business users to recognize and utilize the data.
   • Establishing "business data producers" within the business teams to curate and manage these data products.

3. AI Adoption:

   • Empowering business teams to directly utilize the curated data products in their AI experiments and solutions.
   • Providing guidance, guardrails, and governance to enable business teams to safely experiment with AI, lowering the cost and time to market.
   • Integrating the semantic layer with AI platforms, such as Amazon Bedrock, to allow business users to "chat with the data" and invoke AI capabilities.

MUFG's Journey: Enabling Business Self-Sufficiency

• MUFG's data strategy initially focused on traditional data consolidation and curation, but they pivoted to a more business-centric approach.
• The key trigger for this pivot was the need for faster time-to-market and enabling business users to be self-sufficient in preparing and accessing data.
• MUFG created a mixed model of core data products and functional data products, allowing business teams to build their own derivatives and data solutions.
• They implemented data literacy programs to educate business users on the available data and how to access it.
• The adoption of the semantic layer and data products has been a mixed experience, with some power users being self-sufficient, while others require more guidance and support.

Integrating AI with the Data Ecosystem

• MUFG is leveraging AI capabilities in two ways:
  1. Using AI for data management tasks, such as profiling, lineage, and cataloging, to enhance the data quality and accessibility.
  2. Enabling business users to "chat with the data" using AI-powered conversational interfaces, while maintaining appropriate access controls and governance.
• The semantic layer and data products serve as a centralized and governed platform for feeding data into AI models, improving the quality and insights derived from the AI.
• The flexibility of the semantic layer allows business teams to experiment with various AI agents and solutions without being tied to a specific platform or data model.

Key Takeaways

• Empowering business teams to lead in the AI era requires a strategic approach that bridges the gap between data and AI capabilities.
• The three-layer foundation of data access, semantic layer, and AI adoption enables business teams to become self-sufficient and innovative in their AI initiatives.
• MUFG's journey demonstrates the importance of pivoting from traditional data strategies to a more business-centric approach, fostering collaboration and trust between IT and business teams.
• Integrating AI capabilities with the data ecosystem, both for data management and business user interaction, can further enhance the value and impact of AI-driven solutions.