Abstract
Artificial intelligence is profoundly transforming organisational strategy formation, progressing beyond simple automation of analytical chores to the collaborative development of strategic intent with human participants (Doshi et al., 2025; Atkinson, 2025). This transformation replaces periodic planning cycles with continuous adaptive formation in volatile and uncertain situations (Cristofaro and Giardino, 2025). While conventional models regard technology as a facilitator of preordained decisions, contemporary AI systems, especially large language models (LLMs), operate as active contributors that collaboratively influence strategic direction through iterative human-machine engagement (Trunk et al., 2020; Kim and Marakas, 2024). This article defines "coevolution" as a unique method of strategy development in which LLMs, and associated AI systems interact with managerial judgement throughout the entire strategy cycle, enhancing cognition while maintaining human control over framing, value-based decisions, accountability frameworks, and ethical limitations. Coevolution expands distributed cognition theory (Hutchins, 1995) to include non-human agents with generative capabilities, while utilising New Institutional Theory to frame AI as both an organisational instrument and an institutional influence that shapes legitimacy and generates isomorphic pressures (Rudko et al., 2025).
This study develops a conceptual framework based on a critical literature analysis and abductive synthesis, identifying three primary clusters of mechanisms by which AI affects strategic outcomes. Initially, decision support techniques improve information processing capabilities and facilitate simultaneous scenario assessment, reducing cycle durations while broadening analytical scope (Enholm et al., 2022; Jarrahi et al., 2023). Secondly, insight generating mechanisms utilise pattern recognition and signal extraction abilities to reveal strategic signals contained in intricate, high-dimensional data settings (Li et al., 2021; Han et al., 2025). Third, organised creativity methods enhance combinatorial exploration and systematic option generation, hence decreasing search costs and broadening the solution space accessible to strategists (Gioia et al., 2023; Merilehto and Poudel, 2025). These processes function through temporal process dynamics, in which human judgement and AI capacities mutually adjust during strategy development cycles (Hutchison et al., 2009; Evans and Stanovich, 2013). Eight formal propositions outline the boundary requirements that dictate when these mechanisms result in strategic efficacy instead of superficial adoption that produces minimal organisational benefit.