The Say action: speaks the text that is given in the specified (or a default) voice.
Parameters:
Text: what to say; may include agent attribute values when the name is preceeded by ~ (for example: ~value)
Voice (optional): may specify the voice to use speaking.