02 Retrieval-Augmented Generation (RAG)

02 Retrieval-Augmented Generation (RAG)

This workflow demonstrates how you can visually embed LLM and embedding models into your analytical workflow. It also showcases a RAG implementation, including vector store creation, retrieval, and prompt augmentation.

The workflow accesses data about product reviews from a database, filters the data for analysis, uses GenAI to automatically define sentiment, identify the product, and categorize the problems in the reviews, and finally analyzes the reviews by product, sentiment, and category. It uses RAG with the company's domain-specific product documentation to achieve more accurate results from GenAI.

Access database data for analysis

Connect to an LLM provider

Data analysis

1. Build knowledge base

2. Populate vector store

3. Retrieve relevant chunks

4. Augment prompt

5. Generate responses