Building adaptive audio cues