Machine learning (ML) based Artificial Intelligence (AI) is revolutionizing many areas of modern life, including translating conversations between different languages, identifying people in pictures, and autonomously driving cars. All these advances have created new, more complex, and challenging tasks where even human execution of these events sometimes falters due to lack of specialized training, lack of attention, information overload, or fatigue. Military commanders from the very earliest days have always been faced with the “fog of war”, but in today’s digital environment, not only can we begin to see through fog, but we can begin to glimpse the foundational components of that “fog”.  Similarly, the actions of a Commander in a battlespace are affected by degree of training, attention to detail, and fatigue.   Conducting tactical operations within a Multi-Domain Operation (MDO) environment exponentially increases the complexity of a leader’s Command and Control (C2) with the inclusion of space, cyber-electromagnetic activities (CEMA), and robotic assets, which are likely to drive the OPTEMPO even higher than in the past. Given the pace of operations, the complexities of the Operational Environment (OE) and the speed at which decisions must be made, sometimes minutes, – the Commander and their staff are simply overwhelmed in trying to assess in a timely matter what information is needed to allow the commander to make the best decision possible.  In collaborative work with the Army Research Labs, we have been exploring the application of Deep Reinforcement Learning (DRL) towards generation of automated Decision-Making Assistants. There are many practical uses for such a capability, including exposing novel or optimal Curses of action (COA), identification of timely and unique threat activities that influence friendly plans, and the supervision of autonomous battlefield systems and activities. In this paper we describe the overall approach, describe the project’s status, and provide lessons learned.