Research background and purpose:
Today’s VUCA environment-characterized by increasing volatility, uncertainty, complexity, and ambiguity-is making the limits of both cognitive and analytical capabilities more apparent in organizations’ strategy development processes. The aim of this study is to examine how the strategy development and strategic reasoning capabilities of generative artificial intelligence (GenAI) tools differ across individual VUCA components.

Design/methodology/approach:
The research employs an exploratory design that combines a qualitative scenario method with multi-GenAI-based strategy development and an Artificial Intelligence–Based Evaluation (AI-as-Evaluator) approach. In line with the research objective, four separate scenarios representing each component of VUCA were generated by the generative AI tool called ScholarGPT, and within the context of these scenarios, strategy proposals were developed by the generative AI tools ChatGPT, Gemini, DeepSeek, and Claude. The strategies developed were evaluated by multiple generative AI tools, within the framework of the principles of blindness and non–self-scoring, in terms of the criteria of innovativeness and creativity, feasibility, agility and adaptability, risk level, and market alignment.

Findings:
The findings reveal that although generative AI tools exhibit high performance in “agility and adaptability” across all scenarios, the “complexity” component leads to a systematic decline in the “feasibility” scores of the strategies and a marked disruption in evaluator consistency.

Value added and limitations:
The study expands the VUCA literature from an artificial intelligence perspective and offers a conceptual and methodological framework for research on generative AI-supported strategy development.