Back to Glossary
RAGRetrieval-Augmented Generation
検索拡張生成(けんさくかくちょうせいせい)
IntermediateModels & Architecture
A technique that enhances AI responses by first retrieving relevant information from external sources, then using that context to generate more accurate answers.
Why It Matters
RAG reduces hallucinations and keeps AI responses grounded in up-to-date, verifiable information.
Example in Practice
A customer support bot that searches your company's knowledge base before answering questions.