Derek Feichtinger has an interesting post in which he describes the application of reproducible research and literate programming to management problems. As an example, he considers generating a budget for a pair of related projects. His workflow is to first generate an outline describing his goal and the information he has and to refine that with subheadings as more information becomes available. That provides a history of the project and automatically tracks changes.
I really like this idea. Something to think about.