Artificial intelligence (AI) is increasingly used in health professions education (HPE), while institutional readiness has not kept pace. Faculty report limited AI knowledge, inconsistent policies, and limited guidance for implementing, evaluating, and sustaining AI integration. Because HPE prepares learners for clinical practice, how AI is integrated into education may also affect patient care. This paper introduces RAISE (Readiness, Architecture, Implementation, Stewardship, Embedding), a conceptual framework for evidence-based and evidence-generating AI integration in HPE. The synthesis of RAISE is informed by three implementation science frameworks: Exploration, Preparation, Implementation, Sustainment (EPIS) for process; the Consolidated Framework for Implementation Research (CFIR) for determinants; and Reach, Effectiveness, Adoption, Implementation, Maintenance (RE-AIM) for evaluation. Grounded in a complex adaptive systems perspective, RAISE recognizes that implementation occurs across interacting individual, program, and institutional levels. A relational layer connects these levels and components, while Architecture, Implementation, and Stewardship operate as a coupled cycle in which evaluation informs implementation as it unfolds. Guiding questions support its application across classroom, simulation, and clinical settings. RAISE provides a shared structure for institutions, programs, and faculty to plan AI integration, examine what works, for whom and under what conditions, and determine whether evidence supports continued use. Prospective empirical testing is needed to establish its utility.