LLMs for Strategic Decision-Making in Business Intelligence

This literature review explores the effectiveness of large language models (LLMs) in supporting strategic decision making within the context of business intelligence. It systematically synthesises existing academic literature to examine how LLMs influence the quality, efficiency, and outcomes of strategic decisions in business environments. The review pays particular attention to how LLMs are integrated into existing business intelligence systems, focusing on challenges and documented best practices related to deployment and ongoing use, with a specific emphasis on natural language summarisation and executive dashboards. The objectives of the review include summarising how recent studies have assessed the operational effectiveness of LLMs in decision making, identifying key themes in the literature regarding integration into business intelligence frameworks, reviewing their reported impact on scenario analysis processes, and examining the role of LLMs in supporting knowledge synthesis for decision support systems. Findings from the reviewed literature suggest that LLMs have the potential to support decision making by providing relevant insights that may enhance the accuracy and contextual relevance of information used in strategic processes. Nonetheless, commonly reported challenges include concerns around data privacy, model bias, and the need for high quality training data. Studies also highlight several recommended practices for LLM deployment, including iterative fine-tuning, promoting transparency in AI processes, and implementing user training to ensure effective use. These insights are particularly relevant for the service industry, where timely and informed decision making is critical. The review highlights the growing interest in generative AI within management literature and points to the importance of developing clear strategies for incorporating LLMs into business intelligence systems in a responsible and effective manner.