Well, to be generally speaking about Speech, you can always use Microsoft's Speech APIs.
http://msdn.microsoft.com/en-us/library/system.speech.recognition(v=vs.110).aspx[
^].
But as already said, you need to work around on this project, as it is really very hard to find the specific patterns in the Speech and the audio provided by the Microphone. Grammer, SpeechRecognizer etc these methods are used to find out the patterns of the speaker' voice. You can surf the Microsoft's developer network to find out the required APIs. This question and topic, as it stands, is really
very broad.