AWS re:Invent 2025 - From enterprise data mesh to AI with Amazon SageMaker Unified Studio (IND3322)
Source: Dev.to
Overview
AWS re:Invent 2025 – From enterprise data mesh to AI with Amazon SageMaker Unified Studio (IND3322)
In this session, Rizwan Wangde and Sam Gordon from AWS Professional Services present a framework for building modern data platforms in financial services to support agentic AI. They identify four critical barriers: data silos, trust in data, cross‑organizational governance, and data consumption patterns. SageMaker Unified Studio is introduced as the central solution, demonstrating how data contracts, unified catalogs, and automated lineage tracking address these challenges. The speakers note that 40 % of agentic AI projects are projected to fail before 2027 due to inadequate data foundations, and showcase practical implementations—including Customer 360 use cases—achieving 80 % data discoverability and a 90 % reduction in manual governance processes through a flywheel approach.
Introduction: The Foundation of AI Lies in Data
Hello everyone. Good morning and welcome to the last day of re:Invent. My name is Rizwan Wangde, Senior Cloud Architect with AWS Professional Services, based in Sydney, specializing in data. Over the past few years I have worked closely with large financial services institutions, and today I’ll share what we’ve seen, learned, and delivered as the industry shifts its business and technology mindset.
Hi, I’m Sam Gordon, Senior AI and ML Consultant with AWS Professional Services. I’ve been collaborating with Rizwan on several financial‑services projects, and I’ll cover the second half of our story.
“If you could reason on top of the data you have, what opportunities could you unlock and what problems could you solve?”
— Prompt for reflection on the business impact of a modern data platform.
The Reality Behind Agentic AI: Why Data Foundations Matter
Data is the fuel for high‑quality AI applications. Production‑grade agentic AI systems cannot succeed without a solid data foundation. From my experience building scalable data platforms for large financial institutions, four key elements are essential:
- Breaking data silos
- Creating trust in data
- Establishing cross‑organization governance
- Enabling flexible data consumption patterns for both humans and AI agents.
Current experimentation with agentic AI is abundant, but scaling remains difficult. Gartner predicts that 40 % of agentic AI projects will fail before 2027. Additionally, 52 % of Chief Data Officers believe their data foundations are not ready for AI consumption.
How do we become ready? By addressing the four critical barriers outlined below.
Four Critical Barriers to Modern Data Platforms
1. Breaking Data Silos
Transformation requires more than technology; it demands an organizational muscle to thrive in an AI‑driven future. Data must be democratized, discoverable, and accessible to both humans and AI agents while remaining secure and governed.
2. Trust in Data
Regulatory environments demand traceability. AI agents must be able to reference the source of the data that informed a decision. The quality of autonomous decisions is directly tied to data quality.
3. Cross‑Organization Governance
Agentic AI needs comprehensive access to business data to make sound decisions. Limited data access leads to incomplete context and sub‑optimal outcomes. Governance frameworks must enable controlled, auditable access across the enterprise.
4. Flexible Data Consumption Patterns
Modern organizations consist of multiple business units with distributed data. A robust platform must allow AI agents—and humans—to access data across boundaries efficiently and securely.
The session continues with a deep dive into how SageMaker Unified Studio implements data contracts, unified catalogs, and automated lineage tracking to overcome these barriers.