Robotic systems have been widely adopted for operation in different industrial fields. The design of software for such a system is inherently complex. The scope of such software code includes controlling various system components. As the development process accelerates, they often lead to insufficient documentation, hindering understanding, debugging, and long-term maintenance. This research proposes a novel approach using generative AI to enhance documentation and improve comprehension of robotic system behavior by automatically converting robotic software source code (for ROS-based systems) into Behavior Trees (BTs). BTs provide a hierarchical, graphical representation of robotic tasks and decision-making logic, offering a clear and intuitive visualization of system operation. This automated conversion process bridges the gap between rapidly evolving codebases and comprehensive system documentation, facilitating easier debugging, behavior analysis, and integration of new features within the ROS ecosystem. The proposed approach is experimented with different ROS projects of varying complexity. The derived BTs are also validated by simulating the ROS projects using Gazebo simulator.