Jester," new email"
"Hello there, my name is N.O. and I am seeking your assistance."
Jester, "Remove I am seeking your assistance" -> result: "Hello there, my name is N.O. and"
"I want to take a trip down to the office and chat about something I feel is exciting enough for me and what I think is a great opportunity for you." -> result is appended to the last result.
Jester, "Send email"
The idea is that the name of the software is used to bring the software into a mode of accepting "menu"/"context" commands by voice. The interaction such as when "New Email" is spoken brings up the default voice enabled email ready for voice input. Speaking without starting with "Jester" is handled by the software to be text input to the current program that has the input context. When you stop for period of time a period may be append to the sentence, but you can always say "Jester, continue sentence" and keep going.
I think this just one wishful, but practical way to make voice input useful. I guess I've talked about this prospect before in a private post though :P