Conversational AI includes chatbots, dialogue systems, intelligent personal assistants, and other conversational agents used for dialogue practice, whether spoken or written, for language learning. This chapter explores the evolution of conversational AI in language learning, from its rule-based origins to the advent of large language models (LLMs). It examines the theoretical foundations underpinning the use of chatbots for language learning, including the Interaction Hypothesis, and the influence of conversational AI on key dimensions of language learning, such as foreign language anxiety and willingness to communicate. The chapter synthesizes empirical evidence regarding the effectiveness of both dedicated language learning applications and general-purpose conversational AI systems for autonomous practice and integration into formal teaching contexts. Key findings from these studies are discussed, highlighting improvements in learners’ productive language skills and autonomy, while also drawing attention to concerns related to conversational AI, such as accuracy and ethical use. Moreover, it provides practical recommendations for teachers, addressing guidance strategies, professional development, and tool selection. The chapter further discusses activity design for conversational practice, proposing three key roles: conversation partner, feedback provider, and moderator. Finally, it considers potential challenges in using conversational AI to support language development and suggests ways to mitigate them, emphasizing the importance of thoughtful integration and teacher expertise.