In this module, you’ll dive into how to unlock insights from unstructured data sources using Cortex Search. You’ll learn how to load, parse, and chunk documents, then build a search service that can power both enterprise search and RAG-based chatbots. You’ll explore how hybrid search works (combining keyword and semantic search) and understand how to measure the quality of search results. You’ll also see how to keep your Cortex Search Service fresh by building a pipeline using Tasks and Streams. Last, you’ll build a frontend for the application using Streamlit so you can deploy the app to users. By the end of this module, you’ll have built a working RAG application in a Snowflake Notebook that can answer natural language questions from unstructured documents—something that previously required manual analysis or was simply inaccessible.

This module focuses on turning structured business data—like the kind you find in Snowflake tables—into a natural language interface. You’ll learn how to build applications that convert questions into SQL using Cortex Analyst. That includes defining semantic models to help the LLM understand your data, generating accurate SQL, and returning precise answers. You’ll also learn how to call the Analyst API and build multi-turn conversational applications that interact with your structured datasets. Then, you’ll learn how to take the results from Cortex Analyst, and translate them back to natural language with an LLM. Last, you’ll build a frontend with Streamlit so your application is ready for the business. By the end, you’ll have a working text-to-SQL app that helps business users get the answers they need without writing a single line of SQL themselves.