AI-collaborative method for qualitative sensemaking: an application to entrepreneurial identity

Abstract

Purpose A hybrid approach to human–artificial intelligence (AI) collaboration that is suitable for qualitative research in entrepreneurship is proposed and demonstrated. A two-phase method is presented for analyzing and interpreting qualitative textual data such as may be obtained from respondent interviews, using artificial intelligence tools in collaboration with researcher judgement. Design/methodology/approach In the first phase, natural language processing is used to discover structure and groups within a set of respondents. In the second phase, a large language model is used to propose discussion themes, language usage and implicit sentiments for the groups and to support these propositions by reference to the original data. These proposals are ultimately validated by researcher direct confirmation. Findings The method is demonstrated by analyzing recorded data from eight entrepreneurs interviewed about the influence of their religious beliefs on their practice of entrepreneurship and by developing novel insights from the AI suggestions. This demonstration illustrates the potential improvements in both efficacy and efficiency of research that may be obtained by judicious use of AI, and how concerns about validity and replicability may be addressed. Research limitations/implications The results are used to discuss the value of the method and make recommendations for future researchers. Originality/value The novel approach presented here contributes to emerging dialogues on human–AI collaboration in interpretive research by considering AI not as a passive tool in analysis, but as a potential partner in sensemaking.