Bridging Enterprise Knowledge Management and Natural Language Processing - Integration Framework and a Prototype

Cappel Justus , Chasin Friedrich

Abstract

Despite the rapid advances in AI, most organizations struggle to fully harness the potential that emerging technologies in the realm of Natural Language Processing (NLP) offer. This study deals with the particular challenge of using large language models (LLMs) to enhance the communication of organizational knowledge among employees and with external customers. Traditionally, companies rely on distributing knowledge via websites, internal documents or knowledge management systems, the use of which often proves tedious. In response, this work proposes an integration framework that helps organizations to connect the digital representations of their existing knowledge with LLMs. This integration enables intelligent retrieval and enhances semantic matching of questions and answers based on the knowledge base. Objectives for the framework are derived from insights gathered through interviews with organizations, emphasizing the practical relevance of the proposed solution, and demonstrate the utility of the framework with a prototypical implementation. This research not only represents a contribution to the ongoing research on the organizational applications of LLM-based digital technologies but also outlines the benefits and the limits of current LLM technologies for the enhancement of organizational knowledge management.

Keywords
Knowledge Management; Natural Language Processing; Information Retrieval; Integration Framework; Prototype