Andrea Kropp

You can stand up a RAG pipeline in an afternoon. Getting it to work on real documents — dense tables, scientific figures and irregular layouts — takes months. This session cuts that gap. We'll walk through an end-to-end workflow using LandingAI's Agentic Document Extraction to parse complex, multimodal content and extract structured key-value pairs that break conventional tools. From there, you'll see how parsed output feeds the Snowflake Cortex AI stack — Cortex Search, Cortex Analyst and Cortex Agents — all surfaced through Snowflake CoWork. The entire stack is created with one SQL script and can be versioned for your own use case in about half a day.