After more than a decade, MOOCs have shown great promise in delivering high-quality, affordable education at scale, breaking down geographical and financial barriers. Yet, they continue to face persistently low completion rates, limiting their impact. Contributing factors include limited learner accountability, weak engagement mechanisms, high perceived time demands, and scarce opportunities for meaningful peer interaction. Addressing these challenges in a cost-effective way remains a major hurdle. This lightning talk introduces an innovative approach to improving MOOC engagement and completion. We present a self-service web interface that simplifies scheduling and coordination of small-group, real-time discussion sessions. This tool is complemented by an advanced AI engine that facilitates these sessions, keeping discussions on track and aligned with learning goals. The AI also provides structured, rubric-based assessments of learner engagement.Pilot tests of this approach have yielded promising results. Participation in live sessions has increased significantly, reflecting greater motivation and accountability. Learners report deeper, more meaningful interactions, driven by structured facilitation and active peer-to-peer engagement enabled by the AI. Additionally, the AI generates personalized formative feedback based on individual contributions, enhancing the learning experience and addressing diverse learner needs at scale.By integrating scalable, AI-facilitated human interactions, this approach offers a powerful way to revitalize MOOCs. It makes online learning more interactive, inclusive, and engaging—supporting global learner communities and positioning MOOCs to reclaim their role as a transformative force in education.