Background: Conversational artificial intelligence (AI), including chatbots and large language models (LLM), based agents, is increasingly explored as a digital health intervention in physical rehabilitation. These systems offer potential benefits in patient education, treatment adherence, self-management support, and clinical workflow augmentation. Despite rapid technological advances, a comprehensive synthesis of their applications in physiotherapy, particularly within specific national contexts such as Greece, is lacking. Objective: This scoping review aims to map the existing literature on conversational AI agents in physical rehabilitation, examine their system design and clinical applications, and identify research gaps and future directions relevant to Greek physiotherapy practice.Methods: Following PRISMA-ScR guidelines, peer-reviewed studies, preprints, technical reports, and regulatory documents published between 2010 and 2026 were identified through searches of biomedical, rehabilitation, and computer science databases. Eligible studies involved conversational AI systems applied to physiotherapy or musculoskeletal rehabilitation. Data were extracted on system architecture, interaction modality, clinical use case, reported outcomes, and implementation considerations. Results: The literature shows growing interest in conversational AI for exercise coaching, adherence monitoring, patient education, and triage. Recent studies increasingly employ LLM-based architecture and retrieval-augmented generation to improve response accuracy. However, clinical evidence remains heterogeneous, with limited randomized trials and inconsistent outcome measures. Few studies address linguistic adaptation, cultural appropriateness, or regulatory alignment in Greek healthcare settings. Conclusions: Conversational AI agents represent a promising adjunct to physiotherapy services but remain at an early stage of clinical maturity. Responsible integration into Greek rehabilitation practice will require rigorous evaluation, localization, and regulatory alignment.