inductions from speech to text software