Martin Adam, University of Goettingen

Aaron Baird, Colorado State University

Mari-Klara Stein, Tallinn University of Technology

Alexander Benlian, Technical University of Darmstadt

Varun Grover, University of Arkansas

Monideepa Tarafdar, University of Massachusetts Amherst

Organizations have treated AI – especially generative AI systems such as ChatGPT and Claude – primarily as a productivity aid in individual activities such as drafting, summarizing, coding, searching, and brainstorming. This view is already too narrow. AI is moving from assisting in the production of information artifacts to participating in organizational action. AI agents, or closely related agentic AI applications, are defined as AI-enabled digital artifacts that can interpret goals, perceive relevant states, plan or select actions, interact with humans and other systems, invoke digital tools, and execute or recommend actions with some degree of autonomy. The central shift is therefore from AI that influences how humans think, write, or decide to AI that can be delegated bounded responsibility for moving work forward. The IS community is beginning to articulate a number of research directions pertaining to this shift (e.g., Adam et al., 2025; Grover & Tarafdar, 2026)

This transition matters because it reconfigures core organizational arrangements. The managerial frontier is no longer only how employees use AI, how firms improve prompts, or how models increase individual productivity. It is also what organizations are willing and able to delegate; how autonomy is bounded; how failures are detected before they cascade; and how accountability, contestability, resilience, coordination, control, and human authority are designed into core operational systems. As AI agents become embedded in enterprise systems, digital platforms, customer interfaces, software repositories, procurement systems, analytics pipelines, and approval workflows, agentic organizations emerge. We refer to agentic organizations as sociotechnical arrangements in which AI agents are embedded in organizational routines, systems, and workflows.

IS research is well-positioned to address this transition. The field has strong foundations in managing AI in organizations (Benbya et al., 2021; Berente et al., 2021), IT-enabled organizational change and behavioral shifts in IS use (Grover & Tarafdar, 2026), algorithmic management (Cram et al., 2022), digital platforms and platform governance (Heimburg et al., 2025; Mayer et al., 2025; Wessel et al., 2025), human-AI decision-making and cognitive engagement with AI artefacts (Saffarizadeh et al., 2024; Tarafdar et al., 2026), AI delegation and agentic IS artifacts (Adam et al., 2024; Baird & Maruping, 2021; Queiroz et al., 2024), AI as teammates and agents (Adam et al., 2026; Dennis et al., 2023), and the social consequences of AI (Sabherwal & Grover, 2024). This expertise gives IS scholars a distinctive vantage point because agentic AI is simultaneously a digital artifact, a form of delegated action, a governance problem, a coordination mechanism, and a source of organizational transformation.

At the same time, existing research has not yet sufficiently explained AI agents as embedded participants in organizational action. Agentic organizations require extending this work to settings in which AI agents do not merely provide information to humans, automate predefined routines, or merely support cognition, but participate in organizational processes, generate context-sensitive and probabilistic action possibilities, and reshape expertise, sensemaking, attention, emotions, and professional identity over time. As such, the promise is substantial: Agentic AI can reduce coordination costs, automate routine work, personalize services, support complex problem solving, accelerate software and process innovation, and enable new digital products and business models. It can also reconfigure how value is created and captured when AI agents become boundary resources, digital coworkers, market intermediaries, or operational orchestrators. Yet, the risks are equally consequential: AI agents may act on incomplete or erroneous information, optimize local subgoals rather than organizational objectives, obscure accountability, expose firms to security and compliance failures, and gradually shift human work from substantive judgment to supervision of opaque systems.

This special issue invites research that moves beyond asking whether AI improves performance to examine how and under what conditions AI agents transform the foundations of organizational life. The motivating question is: How can organizations design, govern, and manage human-AI ensembles in which AI agents act with increasing autonomy while preserving accountability, resilience, human judgment, and societal value?

This special issue provides a dedicated forum for IS and related scholars to examine the strategic, organizational, behavioral, and societal implications of agentic organizations and the management of AI agents. We welcome work at all levels of analysis, including the individual, team, workflow, firm, platform, ecosystem, and societal levels. Submissions should make agentic organizations and AI agents central to the theory, empirical setting, design intervention, or analytical model. Papers that treat AI only as a generic automation tool, chatbot, or productivity aid, without engaging with agentic action, do not fit the special issue.

Consistent with the mission of JMIS, submissions should develop theoretically grounded and empirically or analytically rigorous insights into the management, governance, design, use, and organizational consequences of agentic IS. A manuscript fits the special issue when it can clearly identify one or more of the following: the delegated goal, the agent's autonomy, the data or tools the agent can access, the consequential organizational action or workflow move, and the governance/accountability arrangement surrounding that action. We particularly invite work that advances IS theory while offering actionable implications for managers responsible for designing, deploying, governing, or auditing AI agents in organizational settings.

The focus of this special issue is to stimulate theoretically insightful, methodologically rigorous, and field-shaping investigations into agentic organizations and the design, governance, and management of AI agents. Agentic organizations and/or AI agents within organizations should be central to the research question, theorizing, design, or empirical setting. All lenses of inquiry are encouraged. We welcome empirical contributions to the special issue, including (but not limited to) trace data, process data, surveys, experiments, field studies, simulations, qualitative studies, case studies, ethnographies, design science, computational approaches, and longitudinal research. Purely speculative essays without empirical, design, computational, simulation-based, formal, or otherwise systematic analytical grounding will not be considered. Also, consistent with JMIS policies, papers should make a significant novel contribution to the IS field. We are especially interested in papers that develop new constructs, mechanisms, process models, design principles, governance frameworks, measurement instruments, analytical models, or empirically grounded theory that explain how agentic AI changes organizational action, coordination, control, accountability, and value creation. Given the emergent nature of agentic organizations and AI agents, we welcome rigorous studies of pilots, prototypes, simulations, design interventions, and early-stage deployments, provided they yield theoretical insight into organizational action, governance, or design.

Potential topics

• I. Delegated autonomy, control, and accountability in agentic organizations — Agentic AI shifts the managerial challenge from using systems to defining what work can be delegated to systems, under which objectives and constraints, and with which forms of oversight and escalation. Organizations need new delegation contracts that specify permissions, prohibited actions, monitoring obligations, human override rights, audit trails, and remedies. Research can examine how autonomy is calibrated across low- and high-stakes tasks; how accountability is allocated when decisions are produced by humans, agents, vendors, data pipelines, and organizational routines; and when human-in-the-loop arrangements are substantive rather than symbolic. Illustrative research questions: Which tasks, decisions, or workflow moves should be delegated to AI agents, and which should remain under human authority? How should organizations design delegation contracts, decision rights, escalation mechanisms, and override arrangements for agentic AI systems? How should AI (vs. human) agent actions be logged, audited, contested, and remediated when outcomes are produced by distributed human-AI ensembles? Useful constructs include human-on-the-loop, human-in-command, escalation fatigue, rubber-stamping, meaningful supervision, moral crumple zones, and responsibility gaps.
• II. Agentic process redesign and organizational capability building — AI agents may reorganize the logic of work by bridging knowledge and action across enterprise systems, analytics, communication tools, and workflow engines. This raises questions about process redesign, decision rights, organizational structure, AI readiness, agent operations, data governance, workforce upskilling, and value measurement. Research can investigate how firms decide which workflows to make agentic, how they redesign tasks and roles around human-AI ensembles, and how agentic capabilities coevolve with organizational routines, control systems, and culture. We are particularly interested in longitudinal and process-oriented studies that examine how organizations move from isolated AI-agent experiments to institutionalized agentic workflows, and how governance, capabilities, routines, and human expertise evolve across this transition. Illustrative research questions: How do organizations identify which workflows, processes, or decision domains are suitable for agentic redesign? How do AI agents reshape business processes, enterprise systems, organizational routines, coordination mechanisms, and control systems? How should organizations measure the value of agentic workflows beyond individual productivity, including coordination quality, resilience, learning, customer outcomes, and strategic adaptability?
• III. Multi-agent orchestration, interoperability, and digital ecosystems — As organizations deploy multiple specialized agents, the central issue becomes orchestration: how agents divide labor, exchange information, avoid conflicting actions, and optimize overall workflow performance rather than local subgoals. Research can examine orchestrator architectures, agent-to-agent coordination, interoperability standards, agentic platforms, APIs and boundary resources, platform governance, and ecosystem competition. Important questions include when agent swarms create value, when they produce excessive complexity, and how organizations govern AI agents that cross firm boundaries. Illustrative research questions: How should organizations design and govern multi-agent systems, orchestrators, task-routing mechanisms, and agent-to-agent coordination? How can organizations prevent local agent optimization from undermining broader organizational objectives? How should firms govern AI agents that interact across organizational, platform, or ecosystem boundaries?
• IV. Epistemic governance, opacity, contestability, and trust — AI agents generate recommendations, explanations, plans, and actions through probabilistic and often opaque processes. Organizational decision makers, therefore, face a shift from reasoning-based validation to new forms of adequacy assessment, provenance tracking, confidence calibration, benchmarking, certification, and contestability. Research can investigate how users evaluate AI-generated claims when causal pathways are difficult to reconstruct; how professionals make sense of conflicts between human expertise and AI agent output; and how explainability, auditability, and epistemic responsibility are designed into agentic systems. Illustrative research questions: How do humans and organizations evaluate agent-generated recommendations, explanations, plans, and actions when causal pathways are difficult to reconstruct? What roles do provenance records, confidence signals, benchmarks, certification, expert review, explanation, agent identity, authorization, access control, and contestability play in governing trust in AI agents? How do organizations deploying AI agents reshape professional judgment, evidence standards, sensemaking, and epistemic responsibility in organizations?
• V. Cognitive, behavioral, and workforce implications — Agentic AI may change not only what people do but also how they think, feel, and develop expertise. When agents perform first-order analysis, drafting, coding, diagnosis, or planning, human work may shift toward prompting, curating, validating, orchestrating, and intervening. Research can examine cognitive recomposition, cognitive coupling, reliance and dependency, skill formation, professional identity, job redesign, role legitimacy, displacement, reskilling, and human dignity. Especially valuable are studies that trace long-term developmental effects rather than short-term productivity alone. Illustrative research questions: How do AI agents change the nature of human work when employees shift from performing tasks to prompting, curating, validating, orchestrating, and intervening? How does sustained reliance on AI agents affect expertise, skill formation, vigilance, problem framing, emotions, creativity, and professional judgment? How do AI agents reshape professional identity, role legitimacy, workplace dignity, human authority, and the meaning of work?
• VI. Resilience, safety, security, and societal consequences — Because AI capabilities remain uneven across tasks and contexts, agentic organizations must be designed for volatility, drift, tool failures, adversarial inputs, hallucinated actions, privacy violations, and cascading incidents. Research can examine agent observability, incident response, red teaming, fail-safe design, human escalation, access control, authentication, security, and compliance. Beyond the firm, research can investigate how agentic systems redistribute power, knowledge, labor, and opportunity across markets, institutions, and society. Illustrative research questions: How should organizations design agentic workflows to withstand drift, tool failures, adversarial inputs, hallucinated actions, security breaches, and cascading incidents? What forms of observability, monitoring, red teaming, access control, incident response, and recovery are needed for safe and resilient AI agents? How can AI agents be governed to support privacy, fairness, dignity, regulatory compliance, humanistic outcomes, and societal value?

Timeline

• December 15, 2026: Optional extended abstract submission deadline (max. 2 pages, non-binding feedback from guest editors; may be submitted at any time before this date via jmis@uni-goettingen.de for an initial non-binding assessment, without formal peer review)
• February 1, 2027: Full paper submission deadline
• May 1, 2027: First round of reviews completed
• September 1, 2027: Paper revisions due
• November 1, 2027: Second round of reviews completed
• February 1, 2028: Second paper revisions due
• Final decision and publication

Associate editors

• Armin Alizadeh, Technical University of Darmstadt
• Abhijith Annand, University of Arkansas
• Langtao Chen, University of Tulsa
• Christy M. K. Cheung, Hong Kong Baptist University
• W. Alec Cram, University of Waterloo
• Carolina Alves de Lima Salge, University of Georgia
• Thomas Hess, Ludwig Maximilian University of Munich
• Ekaterina Jussupow, Technical University of Darmstadt
• Harris Kyriakou, ESSEC
• Aron Lindberg, Stevens Institute of Tech
• Anne-Sophie Mayer, Ludwig Maximilian University of Munich
• Abhay Mishra, University of Arkansas
• Long The Nguyen, Washington State University
• William Olivera, Florida State University
• Gabriele Piccoli, Louisiana State University
• Magno Queiroz, Florida Atlantic University
• Marta Stelmaszak, University of Massachusetts Amherst
• Timo Sturm, Technical University of Darmstadt
• Scott Thiebes, Tongji University
• Thomas Ware, University of Pittsburgh
• Martin Wiener, Technical University of Dresden
• David Xu, City University of Hong Kong