Quick answer: Retrieval-augmented generation — combine LLMs with private data via embeddings + search.
RAG Architecture (Retrieval-Augmented Generation) combines large language models with your own data sources to generate accurate, contextual responses. Instead of relying solely on an LLM's training data, RAG systems retrieve relevant information from private databases, documents, or knowledge bases in real-time, then feed that context to the model for generation.
This skill lets you build AI applications that stay current with live data—customer support chatbots that access company documentation, enterprise search systems that rank internal documents, medical diagnosis assistants using latest research, or financial advisory tools querying regulatory databases. RAG solves the hallucination problem by grounding LLM outputs in factual, retrievable sources.