The pointy ends of our national defense spears involve some of the most sophisticated technologies in the world. Just behind these pointy ends, however, is a collection of planning and debriefing processes mired in decades-old technologies (e.g., chalkboards, whiteboards, laminated maps) that impede the operational ability to adapt, decide, and execute. This paper describes our research, development, and evaluation efforts for next generation mission planning and debrief. The objective is to simultaneously accelerate processes and improve outcomes. Our approach involves combinations of new technologies that allow human-machine teams to engage in these iterative processes together. The core of our in-house development is a web-based system we call Metis, after the ancient Greek goddess of wisdom, prudence, and deep thought. Metis provides a web interface for people, an API for agents, and a database backend to serve both in a real-time fashion. Using this service-oriented architecture we have begun to develop and integrate an array of agent technologies. These include an air tasking order parser, a reading agent that extracts constraints from mission planning guidance documents, a mission plan generation and validation agent, a planning product development agent, and an automated debrief focus point identification agent. Each of these and their integration is in the prototype stage of development, but already we have begun formative evaluations. These indicate dramatic improvements in workflow and resulting decreases in process completion time. Early comparisons of machine-generated content to human-generated content show that they are often (but certainly not always) comparable. The paper will describe all of these components and the latest evaluation results in detail.