This chapter examines the transition from inherited user habits to intentional prompt literacy in the context of large language models. It begins by identifying the mismatch between search-based interaction patterns and the probabilistic nature of AI systems, which leads to vague inputs and suboptimal outputs. The discussion reframes prompting as structured instruction rather than casual query, emphasizing the need to unlearn habits shaped by search engines and chat interfaces. It introduces prompt literacy as a new form of digital competence grounded in clarity, structure, and strategic articulation. The chapter outlines key linguistic principles, clarity, order, and scope, as essential dimensions for aligning human intent with machine output. It categorizes foundational prompt types, including one-shot, role-based, and instructional prompting, highlighting their respective strengths and limitations. The role of tone, length, and format control is examined as central to shaping usable and context-appropriate responses. Common failure modes such as hallucination, drift, and bloat are analyzed to illustrate risks arising from poorly structured prompts. The chapter further positions prompting as an iterative, dialogic process involving reframing and refinement rather than one-time interaction. Ultimately, it establishes prompt literacy as a disciplined, teachable practice that enables users to guide AI systems with precision, accountability, and adaptability.