Talk to Your Data

Overview

Agents need your own information to answer business questions. That information comes in many shapes, structured data in tables and unstructured data such as text and images, and each shape needs a different treatment. This course goes deep on structured data and semantic layers, covers the principles of unstructured data with text, and you build both hands-on on Databricks.

What you'll cover

• ▸Giving an agent your own information: why tables and text need different treatments
• ▸Why plain text-to-SQL fails: the dream vs reality
• ▸An AI-ready architecture for structured data: data products, metadata and semantic layers
• ▸Unstructured data, the principles with text: chunking, embeddings and retrieval

What you'll be able to do

• ✓Explain why tables and text need different treatments when you give them to an agent
• ✓Make a data product agent-ready with metadata and a semantic model
• ✓Implement an end-to-end talk-to-your-data use case with Metric Views and Genie spaces
• ✓Set up question answering over documents and combine it with your tables in one agent