Traditional retrieval-augmented generation (RAG) systems often produce lengthy and irrelevant content, failing to meet specific user needs.