Development of a RAG-Driven Chatbot for Transparent and Efficient Access to Enterprise Data

In the context of the exponential growth of corporate data, traditional keyword-based information search methods prove insufficiently effective, leading to time costs and reduced employee productivity. This work is dedicated to the development and implementation of an intelligent chatbot based on the Retrieval-Augmented Generation (RAG) architecture to optimize access to corporate data within the Dawiso system. The proposed solution allows users to formulate queries in natural language and receive relevant answers with links to data sources, increasing the speed of access to information and ensuring the transparency of results. The work details the design, technical implementation, optimization (including dialogue management and hybrid search), and effectiveness evaluation of the chatbot. Testing results demonstrate the achievement of the set goals for providing fast and accurate access to information, integration with corporate applications, and increasing user satisfaction. The work confirms the viability of the RAG architecture in a corporate environment and outlines paths for further system development, including the implementation of an agentic approach and integration with graph data structures.