It works by capturing your speech through a built-in microphone, then using AI-powered software (like ChatGPT, Google Voice, or local engines) to convert voice into text or commands in real time.