This chapter develops the theoretical foundations underlying Agentic Artificial Intelligence (AI) by examining the integration of Multi-Agent Systems (MAS), Reinforcement Learning (RL), and language-based reasoning. Multi-agent frameworks are reviewed to establish the principles of decentralised control, interaction, and emergent behaviour. The chapter then extends these foundations by incorporating learning dynamics that enable agents to adapt policies through experience and interaction. A central contribution of this chapter is the examination of communication and coordination mechanisms, with particular emphasis on the role of language as a shared semantic substrate for reasoning and collaboration. By formalising how agents can combine learning, communication, and deliberation, the chapter distinguishes Agentic AI from traditional MAS. The discussion highlights how agentic systems support long-horizon objectives, dynamic goal management, and reflective decision-making, thereby enabling persistent autonomy in complex environments.