Literature review on agentic AI

The rapid advancement of artificial intelligence (AI) has transformed the landscape of manufacturing, ushering in new paradigms such as Industry 4.0 and the emerging vision of Industry 5.0. While traditional AI systems have primarily focused on automation, optimization, and predictive analytics, recent developments point toward agentic AI—systems capable of autonomy, goal-directed behavior, and adaptive interaction with both human operators and other machines. This thesis provides a structured literature review of the foundations, developments, and implications of agentic AI, with particular attention to its relevance in smart manufacturing environments. The review is organized along two trajectories. First, it examines the conceptual underpinnings of agency and agentic behavior in AI, highlighting distinctions between human and artificial agency, the rise of perceived agency, and the design of multi-agent systems. Second, it explores the practical integration of agentic AI within smart manufacturing systems, focusing on areas such as autonomous production planning, predictive maintenance, supply chain coordination, and human–AI collaboration. The analysis is based on a systematic search of peer-reviewed publications, identifying both the opportunities and the challenges of embedding agentic AI in industrial contexts. Findings suggest that agentic AI has the potential to improve efficiency, resilience, and adaptability in manufacturing processes, but its deployment also raises critical questions of trust, transparency, and ethical responsibility. By bridging conceptual insights with real-world applications, this thesis aims to contribute to the understanding of how agentic AI can support the transition toward more sustainable, human-centered, and intelligent manufacturing systems.