Abstract

The rapid growth of enterprise data across digital repositories, cloud platforms, knowledge bases, and document management systems has created significant challenges in information retrieval and knowledge management. Conventional search systems often struggle to deliver contextually relevant responses, whereas standalone Large Language Models (LLMs) can produce inaccurate or misleading outputs due to their limited access to organization-specific knowledge. Retrieval-Augmented Generation (RAG) has emerged as an effective framework that integrates information retrieval with generative artificial intelligence, enabling the generation of more accurate, reliable, and context-aware responses. Despite these advantages, enterprise environments require robust access control mechanisms to protect sensitive information. To address this challenge, this study proposes an Enterprise Retrieval-Augmented Generation (RAG) system integrated with Role-Based Access Control (RBAC), providing secure, intelligent, and role-aware knowledge retrieval within organizational environments. The proposed framework combines vector databases, embedding models, retrieval engines, large language models, and RBAC policies to provide secure, personalized, and context-aware responses. The study discusses system architecture, methodology, implementation strategies, hardware requirements, security considerations, and performance evaluation. The results indicate that integrating RBAC into RAG systems significantly enhances information security while maintaining efficient retrieval performance and high response accuracy. The proposed framework enterprise knowledge management, decision support, and secure AI-powered information access.