Enhancing Academic Search With Agentic Rag and Knowledge Graphs: A Case Study On Econbiz

Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by incorporating external knowledge; however, conventional vector-based retrieval methods often struggle with contextual relevance and multi-hop reasoning. To address these limitations, we propose Agentic RAG with Knowledge Graphs (KG), a framework that integrates structured knowledge representation with AI-driven agentic reasoning to optimize academic search. We implement this approach on EconBiz, a widely used platform for economic research, using a three-stage pipeline comprising: (1) a baseline RAG system, (2) a KG-augmented retrieval module, and (3) an agentic RAG framework featuring autonomous agent-based reasoning. Experimental results show that this approach significantly outperforms baseline models in retrieval precision and semantic synthesis, as evaluated using SBERT similarity, BERTScore F1, and LLM-based metrics. These findings underscore the transformative potential of combining knowledge graphs with agentic reasoning to advance academic search and other knowledge-intensive applications.