Dynamics of human–agentic information systems relationship: a systematic literature review

Purpose A new type of information system (IS) known as agentic information systems (AIS) has emerged in recent years. AIS can autonomously complete complex tasks, adapt dynamically and collaborate with other agents (AI and humans) with minimal human intervention. AIS completely changes the way in which traditional IS used to interact with humans. The purpose of this study is to understand the nuances of human–AIS interaction. Design/methodology/approach This study conducts a systematic literature review using five-stage grounded theory method and analyses 53 papers in the domain of AIS. Findings This study reveals that the most important factors that impact human–AIS interaction are delegation mechanisms, user perception and capabilities and AIS design and capabilities. Agency shifts dynamically between humans and agentic systems as the interaction evolves between both agents. Also, five propositions are provided in this paper. Practical implications Guidelines are given for managers regarding how to delegate task between humans and AIS. This study will help AIS designers in their design of AI tools and also provide help to leaders of organization in overcoming adverse impact of agentic AI. Finally, it also gives direction for policymakers of education to design AI-related courses which will help workforce in working with agentic systems. Originality/value This paper contributes to the literature by synthesizing all the relevant papers regarding human-AIS interaction and gives future research directions in the domain of AIS.