For humans, intuitive inferences are made every now and then. However, it would require reasons for humans to convince others and justify themselves of their inferences or decisions. How can machines better convince humans of their predictions? The key may lie in being right for the right and faithful reasons.