Most of the information financial institutions rely on is still locked in documents such as prospectuses, fund factsheets, financial statements, ESG disclosures, credit agreements and client onboarding packs. The challenge is no longer whether AI can read them. It is how fast you can get trusted data, and what the last 1% of accuracy costs you in time and money. This hands-on workshop explores how firms are using agentic AI to discover, extract, classify and validate data at scale and speed. It covers how to build the quality checks, lineage and human-in-the-loop controls needed to trust the output, how to integrate extracted data with core reference, market and portfolio data, and where automation delivers measurable ROI across risk, ESG reporting and operations.  Participants will work through real use cases and leave with a practical framework for prioritising which document-heavy processes to automate first.