Simulation with standardized patients is essential in health professions education but is often limited by faculty time, scheduling, and cost. Artificial intelligence (AI) offers an opportunity to expand authentic clinical practice while maintaining meaningful feedback and learner reflection. We developed an AI Virtual Standardized Patient (VSP) curriculum that delivers scalable, competency-based clinical encounters to strengthen communication, clinical reasoning, professionalism, and self-directed learning. Faculty created AI-powered VSP cases aligned with physician assistant curricular objectives using structured patient scripts, standardized prompts, and competency-based rubrics. Learners interview virtual patients, develop differential diagnoses, recommend diagnostic testing and management, and communicate care plans. Following each encounter, the AI provides immediate feedback on clinical reasoning, communication, empathy, professionalism, organization, and diagnostic accuracy. Learners can repeat cases, promoting deliberate practice before faculty debriefing or live standardized patient experiences. The curriculum has been implemented across acute care, chronic disease, and difficult conversation scenarios. Early implementation demonstrates high learner engagement, increased opportunities for independent practice, and more consistent formative feedback while reducing reliance on standardized patient availability. AI Virtual Standardized Patients provide a scalable model for competency-based education by extending simulation beyond traditional resource limitations. Through authentic patient interactions, immediate feedback, and repeated practice, AI enhances clinical reasoning, communication, professionalism, and reflective practice while reinforcing the human-centered competencies essential for future healthcare professionals.