How Agentic AI Influences Information Systems Artifact Design: A PRISMA Systematic Literature Review

Agentic artificial intelligence (AI) refers to systems that autonomously decompose high-level goals into sub-tasks, invoke external tools, spawn peer agents, and revise plans in response to environmental feedback. This class of systems constitutes a qualitative departure from the AI that Information Systems (IS) scholarship has previously theorised. This paper examines how agentic AI reshapes IS artifact design through a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) systematic literature review of 64 verified peer-reviewed studies retrieved from six major databases, supplemented by citation tracking to approximate a full 138-study corpus spanning 2018 to 2024. Thematic synthesis reveals six convergent paradigm shifts: adaptive architectures that self-modify at runtime; goal-directed agency that supplants rule-based automation; multi-agent coordination producing emergent system behaviours; fluid human-AI teaming that dissolves classical human-system boundaries; explainability and trust elevated to first-class design properties; and embedded ethical alignment as a structural design layer. We argue that prevailing Design Science Research (DSR) kernel theory requires three substantive extensions, Dynamic Design Completeness, Distributed Artifact Agency, and Continuous Evaluation, and we propose a revised IS artifact ontology distinguishing defined-boundary from open-boundary artifacts together with a fivequestion forward research agenda.