This paper explores how AI coaching can enhance decision-making in cyber risk management, focusing on overcoming cognitive biases like escalation of commitment and groupthink. These biases often hinder effective risk management by causing teams to stick to failing strategies and suppress dissenting opinions. Integrating AI into Rasmussen’s Risk Management Decision Model offers a data-driven, unbiased approach to improve decision-making. The authors used Network Oriented Modeling to provide objective insights throughout the risk management process, from identification to mitigation. AI Coaching helps counteract biases such as overconfidence and confirmation bias, leading to more rational and informed decisions. The study emphasizes the importance of adaptive, knowledge-based reasoning, facilitated by AI, to handle complex, evolving cyber threats. The paper includes simulation experiments based on network oriented modeling demonstrating how AI coaching can effectively manage cyber-attack scenarios by mitigating groupthink and enhancing strategic planning. Different scenarios show the impact of AI activation timing and team familiarity with knowledge-based approaches on the effectiveness of the response. Overall, the research highlights the potential of AI to improve cyber risk management by providing continuous, unbiased support, reducing vulnerability to cyber threats, and fostering a more adaptive and resilient security posture. Future research should focus on the long-term implications of AI in decision-making, human-AI collaboration, and ethical considerations.