The landscape of large language models (LLMs) has expanded, resulting in a need for clarity on retrieval augmented generation (RAG) concepts and methods. RAG systems involve retrieving information from data sources to generate LLM output, with various approaches addressing challenges such as query transformations, expansion, re-writing, compression, routing, construction, text-to-SQL, text-to-Cypher, and indexing. These methods enable robust retrieval, efficient querying of structured and unstructured data, and improved performance through techniques like chunk size tuning and document embedding strategies. Post-processing steps also address document combination and ranking, with future plans focusing on open-source models and benchmarks to evaluate RAG approaches.