Introduction: From Assistance to Autonomy
Information systems (IS) research currently finds itself in a phase of conceptual reorientation: generative AI systems are no longer understood primarily as reactive assistance tools but increasingly as agentic information systems (Agentic IS) that independently pursue goals, initiate actions, and assume responsibility for tasks characterized by high uncertainty (Baird & Maruping, 2021). Agentic systems are thus no longer passive tools but artifacts capable of independently assuming rights and obligations in pursuit of goal attainment (Holldack et al., 2026). For IS research, this opens up a field of inquiry at the intersection of Human-Computer Interaction (HCI), organizational design, and intelligent information systems.
Theoretical Positioning and Conceptual Foundations
The growing prevalence of agentic systems calls for an advancement of established theories of socio-technical systems design. Whereas classical approaches in IS use research grant the human actor analytical primacy and treat technology as a passive tool, Baird and Maruping (2021) argue that this assumption must be revisited for a new generation of "agentic" IS artifacts. To this end, they introduce the concept of delegation as a theoretical lens for explaining the relationship between humans and agentic IS artifacts along the dimensions of endowments, preferences, and roles, as well as the mechanisms of appraisal, distribution, and coordination (Baird & Maruping, 2021).
In addition, the Task-Technology Fit (TTF) model provides an established framework for the fit between task requirements and technological characteristics (Goodhue & Thompson, 1995). For agentic architectures, this model needs to be extended along the following dimensions:
- Degree of autonomy: the extent of independent decision-making and action initiation (Baird & Maruping, 2021).
- Intentionality: the capacity for goal-directed planning across multiple sequential actions (Holldack et al., 2026).
- Fault tolerance and resilience: the ability to manage uncertainty within dynamic environments.
- Transparency: the traceability of decision logics and action pathways.
Methodological Approaches and Empirical Research
Studying agentic AI requires methodological designs that go beyond pure performance metrics and account for socio-technical effects. Multi-method research approaches are particularly well-suited to this task:
- Design Science Research (DSR): the development and evaluation of artifacts in the form of AI agents or multi-agent systems, differentiated by instantiations, models, and methods, using both ex-ante and ex-post evaluation designs (Hevner et al., 2004).
- Theory-driven delegation research: applying the delegation framework to model testable hypotheses about appraisal, distribution, and coordination mechanisms (Baird & Maruping, 2021).
- Qualitative case studies: examining organizational transformation processes and socio-technical interactions during the adoption of agentic systems (Yin, 2014).
Challenges and Key Research Areas
A central tension concerns the traceability of autonomous decision and action chains. Autonomous AI agents violate core assumptions of classical accountability attribution, since actions can no longer be unambiguously traced back to individual human decisions (Rahwan et al., 2019). Accountability must therefore increasingly be situated at the level of organizational governance architecture, which determines which goals an agent pursues and to what extent decision-making authority is delegated (Rahwan et al., 2019).
This gives rise to several key research areas:
- Agentic governance: developing frameworks for steering, controlling, and auditing autonomous systems.
- Delegation theory: empirically testing the delegation mechanisms proposed by Baird and Maruping (2021) in real organizational contexts.
- Shifting roles and competencies: examining the effects of agentic delegation on decision-making structures and managerial accountability.
- Process integration: embedding agentic IS within existing process landscapes.
Implications for Research
Agentic AI does not represent an incremental advancement but requires a fundamental revision of IS use theory, as reflected in the MIS Quarterly Executive Special Issue on Agentic AI (MIS Quarterly Executive, 2026). For IS research, this creates a pressing need to empirically validate models of delegation, governance, and trust. In particular, the interface between human oversight and machine autonomy remains a central focus of future research.
References
- Baird, A., & Maruping, L. M. (2021). The next generation of research on IS use: A theoretical framework of delegation to and from agentic IS artifacts. MIS Quarterly, 45(1), 315–341. https://doi.org/10.25300/MISQ/2021/15882
- Holldack, F., et al. (2026). Agentic information systems. Electronic Markets 36, 5. https://doi.org/10.1007/s12525-025-00861-0
- Rahwan, I., et al. (2019). Machine behaviour. Nature, 568(7753), 477–486.
- Goodhue, D. L., & Thompson, R. L. (1995). Task-technology fit and individual performance. MIS Quarterly, 19(2), 213–236. https://doi.org/10.2307/249689
- Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems research. MIS Quarterly, 28(1), 75–105. https://doi.org/10.2307/25148625
- Yin, R. K. (2014). Case study research: Design and methods (5th ed.). Sage.
- MIS Quarterly Executive. (2026). Special issue call for papers: Are organizations ready for autonomous AI agents? https://lnkd.in/dx-wesDZ, accessed August 1, 2026.