As communication service providers progress toward Autonomous Network Level 4 (AN L4), change management becomes a critical bottleneck. While AI agents are increasingly used for monitoring and optimization, network changes—such as configuration updates, capacity adjustments, and service modifications—remain largely manual, risk-averse, and slow. This limits operational efficiency and prevents CSPs from fully realizing agile, intent-driven operating models. The Managing Change in AN L4 Catalyst addresses this gap by extending the agentic AI framework to include intelligent, governed change management. The project explores how human and machine collaboration can be orchestrated to safely plan, assess, execute, and learn from network changes in an AN L4 environment, shifting change operations from static procedures to adaptive, AI-assisted workflows. The solution introduces an agent-driven change orchestration approach, leveraging chain-of-thought reasoning to evaluate change intent, assess risk and impact, recommend actions, and coordinate execution across domains. Rather than replacing human oversight, the framework enables progressive delegation—allowing AI agents to manage low-risk, repeatable changes autonomously while keeping humans in the loop for high-impact or exceptional scenarios. Aligned with TM Forum Autonomous Network principles, the Catalyst defines a structured methodology and performance metrics for AI-assisted change management. Success is measured across four dimensions: execution reliability, safety and compliance, operational efficiency and velocity, and system intelligence and learning. These metrics provide CSPs with a clear, measurable path to scaling automated change while maintaining trust and control. By transforming change management into an intelligent, collaborative process, the Catalyst enables CSPs to increase operational efficiency, reduce risk, and accelerate the transition to truly autonomous, agile network operations at AN L4.