Abstract This thesis presents the design and implementation of GAIA Chat, a conversational, agentic layer built on top of the GAIA Meta-Platform to provide a human centered interface for federated energy and water consumption data. The system supports multi agent orchestration and addresses the inherent limitations of traditional dashboard based and static reporting solutions. Conventional approaches to data exploration rely on the assumption that stakeholders possess the specialized technical skills required to navigate complex query mechanisms, an assumption that no longer scales in today’s increasingly fragmented and heterogeneous IoT environments. GAIA Chat mitigates these barriers through an agentic framework that leverages Large Language Models to provide autonomous intent interpretation, dynamic module invocation, and comprehensive visual synthesis of cross-domain data.