As the number of robotic systems on the battlefield increases, the number of operators grows with it, leading to significant cost burden. Autonomous robots are already capable of task execution with limited supervision, and the capabilities of autonomous robots continue to advance rapidly. Because these autonomous systems have the ability to assist and augment human soldiers, commanders need advanced methods for assigning tasks to the systems, monitoring their status and using them to achieve desirable results. Mission Command for Autonomous Systems (MCAS) aims to enable natural interaction between commanders and their autonomous assets without requiring dedicated operators or significantly increasing the commanders’ cognitive burden. This paper discusses the approach, design and challenges of MCAS and present opportunities for future collaboration with industry and academia.