Why Don't You Do Something About It? Outlining Connections between AI Explanations and User Actions