Recall

Recall measures the proportion of relevant items successfully retrieved in information retrieval systems. It is calculated as relevant_retrieved / total_relevant.

Case Study: RAG Recall Enhancement

• Improvement: A client project increased RAG recall from 50-60% to >95% by leveraging large-language-models for advanced data indexing and structured query generation (see 2026 04 14 Improving RAG accuracy for retrieval).
• Initial Setup: Classic RAG application for internal customer service chatbot, retrieving data from multiple customer databases and document repositories.
• Key Techniques:
  • LLM-powered data indexing (beyond simple vector search)
  • Structured query generation to align with user intent

Source Notes

• 2026-04-23: https://www.youtube.com/watch?v=6vG_amAshTk Video by Adam Lucek This video provides a detailed introduction to Knowledge Graph RAG (Retrieval Augmented Generation), contrasting it with traditional vector database retrieval and demonstrating its implementation using Microsoft’s (Knowledge Graph or Vector Database RAG comparison)