Use when building LLM applications that need to answer questions about proprietary, recent, or domain-specific knowledge that is not in the model's training data
