The integration of Artificial Intelligence (AI) into education has heralded new paradigms for language learning, with conversational AI chatbots emerging as potent tools for fostering learner autonomy. Therefore, this article explores how conversational AI chatbots support and reinforce Self-Regulated Learning (SRL) in language acquisition. It examines how conversational AI chatbots scaffold self-regulated language learning (SRLL) through personalised, adaptive, and metacognitive support, drawing on Zimmerman's cyclic model of self-regulated learning (SRL) and Winne and Hadwin's COPES framework. It synthesises empirical insights, particularly from Abdallah (2024) on developing EFL student teachers' skills via an AI-chatbot SRL model, with established SRL theoretical frameworks (e.g., Zimmerman, Winne & Hadwin) and recent literature. It also examines mechanisms by which chatbots enhance key SRL cyclical phases--forethought (goal-setting, planning), performance (monitoring, strategy use), and self-reflection (evaluation, adaptation)--through features like immediate personalised feedback, adaptive scaffolding, and opportunities for metacognitive engagement. Pedagogical implications of Abdallah's (2024) four-phase model for designing effective chatbot-assisted language learning are highlighted, focusing on fostering motivation, engagement, and learner autonomy. It proposes five principles for effective chatbot implementation: strategic prompting, metacognitive scaffolding, affective-aware feedback, data-driven personalisation, and ethical transparency. Hence, the article contributes a conceptual roadmap for employing AI to transform language education into a student-centred, lifelong learning paradigm. The transformative potential of AI chatbots is discussed alongside challenges in ethical design, contextual adaptability, and avenues for future research to optimise their role in cultivating self-regulated language learners.