Better collection process

I realize this may be asking for a complete re-architecture, but I find the process of the active discovery / multi-instance collection to be highly inefficient, at least when it comes to scripting...