Unifying Data for the Future: The World Bank’s Path to an AI-ready Lakehouse

This session explores how the World Bank’s Finance & Treasury organization is modernizing its data foundation to support mission-critical analytics and AI. You’ll learn how the team consolidated 70%+ of finance data into a single Azure-based lakehouse, standardized on Parquet and Dremio, and eliminated redundant copies to create a single source of truth for treasury, trading, and risk. We’ll walk through the bank’s architecture—from secure onboarding and governed access via Entra ID to a centralized semantic layer in Dremio that powers self-service analytics for finance users while preserving strict controls and row-level security.

You’ll also see how this AI-ready lakehouse is enabling real business impact, including an AI-driven workflow that extracts and validates bond term sheet data, cutting trade setup from 6–8 hours down to minutes with human-in-the-loop review and 95% confidence. The session closes with lessons learned, the roadmap toward unified cataloging and agentic AI, and how open, lakehouse-native patterns are helping one of the world’s leading financial institutions safely accelerate automation and decision-making.