Agent-mediated Knowledge Recovery: The Case of Turnover-induced Knowledge Loss in a Knowledge-intensive Institution

Abstract

The retention of tacit knowledge within organizations, particularly during employee transitions, poses a significant challenge in business settings. This problem is often exacerbated by the underutilization of technology to structure and retain this knowledge. To address this issue, this study proposes an agent-mediated model for knowledge recovery aimed at minimizing knowledge loss by identifying essential information in knowledge repositories and transforming unstructured data into structured format. This study adopts a qualitative case study methodology, employing semi-structured interviews and observational techniques for data collection. Findings, as elucidated through Process-People-Technology (PPT) analysis, highlight the value of understanding routine human tasks in creating rulebased process flows. The proposed CIDM Model (Connect-Identify-Decide-Method) entails of agents, equipped with natural language processing capabilities, which may discern, comprehend, and act on unstructured messages. Additionally, the model enables these agents to identify the roles and responsibilities of message senders and recipients, leading to a more informed decision-making process concerning the use of obtained information. The study highlights the importance of a solution for a more structured, efficient means of retaining organizational knowledge, potentially revolutionizing current management practices.