AI-Augmented Strategic Decision-Making Under Time Constraints: An Experimental Study on Mental Representations and Strategic Foresight.

This article examines how time constraints and the use of large language models (LLMs) jointly influence mental representations and strategic foresight—the ability to predict strategic outcomes—in a startup evaluation task involving 348 participants with business strategy experience. Using a 2×2 experimental design, the study finds that both time constraints and LLM use significantly alter the breadth, depth, and consensus of mental representations: time constraints reduce breadth and increase consensus without affecting depth, while LLM use increases breadth, decreases depth, and reduces consensus. Despite these shifts in mental representations, neither time constraints nor LLM use significantly improve or impair strategic foresight accuracy. Additional analyses reveal that LLM use increases information overload and reduces psychological ownership of decisions, suggesting cognitive bottlenecks that may limit the effectiveness of LLMs in strategic decision-making under time pressure. The findings highlight the need for further research on how individual cognitive processes and contextual factors interact with LLM-augmented decision-making to enhance strategic foresight.