STRUCTURING STRATEGIC DECISIONS WITH LARGE-SCALE LANGUAGE MODELS: AN APPLICATION OF SCA IN SUSTAINABLE FASHION RETAIL

This study proposes an artifact to support the organization and understanding of unstructured textual data as inputs to the Strategic Choice Approach (SCA), through integration with large-scale language models (LLMs). Based on the Design Science Research (DSR) approach, the artifact was applied to a set of interviews with 12 experts from the sustainable fashion retail ecosystem in Brazil. The results demonstrate that LLMs, under the ChatGPT® interface, were able to identify recurring patterns and systematize qualitative perceptions in decision areas, areas of comparison, and areas of uncertainty, aligned with the logic of SCA. It was observed that the digitalization of the chain emerges as a structuring element of strategic decisions, while factors such as consumer behavior, economic viability and institutional coordination are relevant sources of uncertainty. The study contributes by demonstrating the potential scalability of problem structuring methods in the face of integration with LLMs.