APPLICATION OF LLM AND RAG MODELS IN IMPROVING ENTERPRISE ELECTRONIC DOCUMENT MANAGEMENT SYSTEMS

Abstract

This article discusses the theoretical aspects of the use of artificial intelligence technologies in improving the electronic document management systems of the enterprise. In particular, modern trends in the field of natural language processing, the possibilities of Large language Model and Retrieval-Augmented Generation models in working with documentary information are analyzed. The limitations of traditional electronic document management systems are shown and the importance of improving them through semantic search and ways of presenting context-based information is substantiated. The RAG model is described as an effective approach that increases the accuracy of responses by combining the generation capabilities of language models with specific documentary data. In addition, the impact of the use of these technologies on the efficiency of Information Retrieval, interaction with the user and information security requirements will be discussed. The results of the study allow us to consider the LLM and RAG models as a promising direction in the intellectualization of electronic document management systems of the enterprise, and also determine the potential for their application in practical systems in the future.