Human-Agent Collaboration in Decision-Making: A Systematic Review of Agentic AI in Augmenting Human Expertise in Healthcare, Finance, and Governance

This systematic review examines how agentic Artificial Intelligence (AI)—autonomous, goal-driven systems—enhances human decision-making across healthcare, finance, and governance domains. Conducted following PRISMA 2020 guidelines, the study analysed 17 peer-reviewed articles published between 2015 and 2024, selected from databases including IEEE Xplore, Scopus, SpringerLink, and PubMed. Agentic AI was found to enhance human expertise through predictive accuracy, real-time responsiveness, and context-sensitive ethical reasoning tailored to specific sectors. The findings reveal that in healthcare, agentic AI supports diagnostics, treatment planning, and hospital operations by synthesising vast datasets and providing ethical, context-sensitive recommendations. In finance, AI agents automate credit analysis, investment strategies, and fraud detection, often outperforming traditional statistical tools. Governance applications include smart city platforms, policy simulation, and civic engagement systems that promote transparency and citizen-centred feedback. Despite these advancements, the study also highlights major concerns regarding explainability, ethical accountability, and regulatory oversight, especially in non-Western contexts. The paper concludes that while agentic AI holds transformative potential, its responsible integration requires interdisciplinary collaboration, transparent system design, and adaptive governance. Future research should focus on real-world deployment studies, the development of inclusive regulatory frameworks, and culturally aware AI design to avoid digital inequity. This review serves as a foundational guide for stakeholders seeking to navigate the ethical, legal, and technical complexities of human-AI collaboration in decision-making.