For projects I am familiar with, usually you have to propose actual deliverables, and there are project reviews and reports that have to be delivered to program

For projects I am familiar with, usually you have to propose actual deliverables, and there are project reviews and reports that have to be delivered to program managers. But given the nature of research, it is kind of unpredictable. One trick is to propose research that you've already completed or at least gotten into a prototype phase.