Retrieval Augmented Generation

Overview

In this chapter, you’ll extend Pronto’s agent with retrieval augmented generation (RAG) so it can answer questions using unstructured support documentation (PDFs).

What you’ll do

• Ingest Pronto support PDFs into an Einstein Data Library
• Configure a Search Retriever that returns relevant chunks
• Ground a Prompt Builder Flex template with Einstein Search
• Use that grounded prompt from the Pronto Service Agent in Agentforce Builder

Continue with the exercises

• Exercise 1: Ingest Unstructured Data
• Exercise 2: Ground a Prompt with Einstein Search
• Exercise 3: Power Agents with Unstructured Data