Text embeddings excel at encoding unstructured text but struggle with structured data operations like filtering, sorting, and aggregating. To overcome these limitations, knowledge graphs and structured tools can be used to provide precision and flexibility in RAG applications. The proposed solution involves using tools designed for structured data, such as Cypher queries, to address complex user queries that require metadata filtering, sorting, and aggregation. By combining structured data approaches with unstructured text search techniques, more accurate and relevant responses can be delivered, enhancing the user experience in RAG applications.