Artificial intelligence systems are increasingly involved in high-stakes decision-making, such as healthcare, financial, and educational determinations. Many have called for explainable AI (XAI), which are AI systems that provide human-understandable explanations for their reasoning or responses. Through explanations, developers and researchers aim to create AI systems that allow human oversight and improved decision-making while fostering trust. While XAI has traditionally focused on developers in their pursuit for improved system performance and fairness, I consider the question of how XAI systems can support non-technical end-users. A core assumption of XAI is that explanations are actionable. That is, they change what users know, enabling them to act using the AI. What makes explanations actionable for end-users? What can happen if we ignore actionability? What human factor dimensions might we be overlooking in our AI designs? In this talk, I will attempt to focalize explainable AI on these questions and give evidence on how asking these questions can improve outcomes for users of AI systems.