University of Qom Journal of Data Analytics and Intelligent Decision-making 3115-8161 2 2 2026 06 30 Artificial neural network evaluating nonlinear evolutionary effect of cognitive and risk complexity in Ambidextrous strategy renewal 10 30 4583 10.22091/jdaid.2026.15918.1050 EN Mohammad Molani Aghdam Corresponding Author, Associate Prof, Innovation and Management Research Center, Ayatollah Amoli Branch, Islamic Azad University, Amol, Iran. Email: mohammadmolani@gmail.com Iliya Ardeshir Larijani MSc, Department of Management, Faculty of Management and Economics, Islamic Azad University, Science and Research Branch, Tehran, Iran. Email: iliyaardeshirlarijani@gmail.com 0009-0009-2513-3743 Journal Article 2026 04 01 This research examines the micro-foundations of individual ambidexterity as a managerial dynamic capability for addressing the tension associated with pursuing dual strategies. It empirically investigates the coevolutionary intertwining effects of different levels of integrative complexity (IC) and business complexity to identify the optimal composition of cognition–risk interaction that enhances ambidexterity performance. A multidisciplinary approach was adopted to address dualities and ambidexterity through a managerial cognition lens. The application of artificial neural networks has helped to gain a deeper understanding of the role of the cognitive dimension related to dualities. The results clarify that risk and business complexity can have both negative and positive effects. Specifically, in low-risk settings, high levels of IC are not capable of inducing higher ambidexterity performance, suggesting that risk should stimulate the cognitive complexity of CEOs. Moreover, the findings reveal that a high degree of cognitive complexity cannot be viewed as the sole dominant factor in successful strategy renewal, as it is exogenously fueled. An investigation of different scenarios indicates that, at higher levels of risk, differentiation-dominant IC, encompassing a perceptual decision-making style can enhance ambidexterity despite limited integration. In contrast, in low-risk settings, integration supported by analogical thinking may be more effective. https://jdaid.qom.ac.ir/article_4583_a5b233cdf1c71219d6a885e9da8644f2.pdf