Agentic AI in the next frontier of model-based software engineering: the arrival of AI-hyper-agile software engineering methods?

Model-Based Software Engineering (MBSE) has matured into a rich ecosystem of methods and tools that support the systematic design, analysis, and evolution of complex systems.Model-Based Systems Engineering (MBSysE) is now following a similar trajectory, building on the advances of MBSE.Over the past decades, we have developed ways to encode our understanding in models, to formalize designs, and to automate transformation and validation steps.In contrast, the recently emerged Agentic AI enables the development of autonomous AI systems that can perceive, reason, plan, and act independently.Such systems can pursue complex goals with minimal human oversight, moving beyond simple command-response interactions to proactively leverage external tools.In this setting, AI agents exhibit agency: They demonstrate goal-driven behavior and can execute multi-step tasks.In software, and more broadly, systems development projects, such AI agents can be viewed as junior, or even senior, developers capable of executing substantial tasks.This creates both a challenge and an unprecedented opportunity for MBSE.At present, it remains unclear whether agentic AI can reliably translate informal, high-level requirements into correctly designed, implemented, and validated systems.The end-to-end process is long, and the risk of hallucinations or other incorrect actions remains significant.