Technological advancements, such as agentic artificial intelligence (AI) and increasingly complex workplace technology, challenge ontological assumptions underlying Information Systems (IS) theorizing, referring to metaphysical beliefs about what entities, properties, and relations constitute sociotechnical reality. As this reality evolves, these assumptions must be reexamined to avoid limiting the explanatory power, theoretical development, and practical relevance of IS research. While IS research acknowledges the existence of different ontological stances (ontological pluralism), greater attention to when and how to productively engage with ontological alternatives could enhance both theoretical and empirical contributions. The dissertation's primary contribution is establishing ‘ontological reflexivity’ as a practice of disciplined pragmatism involving: (1) sensitivity to ontological pluralism and awareness that one’s ontological stance represents only one of multiple defensible possibilities, (2) articulation of ontological assumptions, and (3) criterion-based evaluation and justification of ontological alternatives. This extends pragmatism in IS research from a predominantly epistemological framework to include disciplined ontological examination. The dissertation contributes practically by demonstrating how ontological reflexivity can be practiced, including reflection on how to balance ontological exploration against the risks of uncritical eclecticism and ontological relativism—both of which can undermine scientific rigor—by making implicit ontological assumptions explicit, assessable, and defensible. Its individual articles contribute discourse-specific knowledge, including comprehensive analysis of ‘agency’ conceptualizations in IS research, introduction of ‘foundation models’ as a new AI paradigm, development of an affordance-based AI agent governance framework, and a resource-sensitive perspective on workaround management. The research comprises two complementary streams across six articles which employ diverse methods including systematic literature reviews, problematization, conceptual analysis, and qualitative case studies with ethnographic observation and interviews. Research Stream 1 establishes the theoretical necessity of ontological reflexivity in light of technological progress, at the examples of conceptualizing increasingly agentic IS and AI foundation models. Moreover, it shows how ontological reflexivity can advance IS theory and field-level contributions by articulating a relational perspective on AI risks to develop an AI agent governance framework grounded in affordance-theory. Research Stream 2 examines ontological reflexivity in practice, demonstrating how it can advance empirical research and practical insights. It presents two qualitative case studies of a critical care information system, showing how ontological assumptions inform researchers’ data interpretation and theoretical connections, and how ontological assumptions embedded in IS artifact design materialize in user experience. By establishing the necessity and value of reconsidering ontological assumptions for IS research, demonstrating how this can be accomplished, and positioning ontological reflexivity within the field's pragmatist tradition, this dissertation addresses a critical gap as technological evolution disrupts established understandings of sociotechnical phenomena. It contributes to the field's capacity to respond to technological change and to positioning IS research as a valuable contributor to emerging cross-disciplinary discourse where sociotechnical perspectives are urgently needed, such as in AI governance.
