Case Study on AI-Based Analysis of Transcribed Mediation Sessions

Abstract

The paper begins by examining the current state of mediation research, which primarily focuses on secondary empirical data obtained through interviews. The authors propose an AI-supported approach—a novel one for professional research—which is based on transcribed mediation sessions. In the context of an original investigation into the actual mediation process, the term ‘analysed primary data’ will be used here. In addition to the conceptual derivation of the analyses carried out, selected results (including a comparison of AI models) are discussed in detail. The case study was carried out as part of the TAHAI (TrustAdHocAI) research project. The aim was to test off-the-shelf AI APIs and low-code development approaches with a view to democratising AI. In line with this framework, the aim was to support the close collaboration of various experts within the context of an agile project organisation.