this is easier for us though because voice ai isn't our core feature so