AI-based diagnosis of the continuity of research thought: A case study from the perspective of innovation and management accounting

<p style="text-align:justify"><span style="font-size:11pt"><span style="background:white"><span style="line-height:105%"><span style="font-family:"Times New Roman",serif"><span style="letter-spacing:-0.1pt"><b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">Purpose</span></span></span></span></span></b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">: This study presents an original test of whether generative artificial intelligence (AI) can detect the internal conceptual coherence within a set of papers produced over several years. The analysis is based on five papers on innovation in the insurance sector, evaluated using M365 Copilot (GPT-5 model).</span></span></span></span></span></span></span></span></span></span></p><p style="text-align:justify"><span style="font-size:11pt"><span style="background:white"><span style="line-height:105%"><span style="font-family:"Times New Roman",serif"><span style="letter-spacing:-0.1pt"><b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">Methodology/approach</span></span></span></span></span></b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">: The full texts of five conference papers, co-authored by the author of this study, were provided to the AI tool, and it was asked to reconstruct the conceptual development and main research direction across them.</span></span></span></span></span></span></span></span></span></span></p><p style="text-align:justify"><span style="font-size:11pt"><span style="background:white"><span style="line-height:105%"><span style="font-family:"Times New Roman",serif"><span style="letter-spacing:-0.1pt"><b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">Findings</span></span></span></span></span></b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">: The AI tool identified both the chronological order and thematic development of the articles, showing a consistent progression from general reflections on risk toward </span></span></span></span></span><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black">a more structured, taxonomic model. However, result quality depends strongly on prompt</span></span></span></span><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt"> clarity and user experience.</span></span></span></span></span></span></span></span></span></span></p><p style="text-align:justify"><span style="font-size:11pt"><span style="background:white"><span style="line-height:105%"><span style="font-family:"Times New Roman",serif"><span style="letter-spacing:-0.1pt"><b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">Research limitations/implications</span></span></span></span></span></b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">: The small, specific sample limits generalisability, and the analysis is bound by the content provided and the model's current capabilities. AI can support self-assessment and exploration of research development but should not replace peer review or formal evaluation.</span></span></span></span></span></span></span></span></span></span></p><p style="text-align:justify"><span style="font-size:11pt"><span style="background:white"><span style="line-height:105%"><span style="font-family:"Times New Roman",serif"><span style="letter-spacing:-0.1pt"><b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">Originality/value</span></span></span></span></span></b><span lang="EN-GB" style="font-size:9.0pt"><span style="line-height:105%"><span style="font-family:"Century Schoolbook",serif"><span style="color:black"><span style="letter-spacing:0pt">: A literature check conducted by the tool suggests that this use of AI to analyse long-term conceptual coherence in an author's work is relatively new. The study contributes to the literature on AI in research methodology by applying AI to assess the development of a research agenda over time, rather than single publications. It shows that LLMs can connect dispersed ideas, highlight practical implications, and suggest future research directions, offering a useful tool for researchers reviewing the evolution of their own work.</span></span></span></span></span></span></span></span></span></span></p>