Toward Agentic AI: Generative Information Retrieval Inspired Intelligent Communications and Networking

The increasing complexity and scale of modern telecommunications networks demand intelligent automation to enhance efficiency, adaptability, and resilience. Agentic AI has emerged as a pivotal paradigm for intelligent networking, enabling AI-driven agents to perceive, reason, and act within dynamic environments. Effective decision-making in telecom applications, such as planning and resource management, requires integrating retrieval mechanisms that support multi-hop reasoning and compliance with evolving 3GPP standards. This article presents a forward-looking perspective on generative information retrieval-inspired communications and networking, highlighting the role of knowledge acquisition and reasoning in agentic AI. We first review retrieval strategies, including traditional, hybrid, semantic, knowledge-based, and agentic contextual retrieval, and analyze their applicability in telecom scenarios. We then survey their use in communication systems, followed by the introduction of an agentic contextual retrieval framework that enhances telecom-specific planning through multi-source retrieval, structured reasoning, and self-reflective validation. Experimental results show that our approach improves answer accuracy, explanation consistency, and retrieval efficiency over baseline methods. Finally, we outline future research directions.