From Tools to Actors: How Agentic AI Reshapes Organizational Learning Dynamics
Type
conference paper
Date Issued
2026-12
Author(s)
Abstract
Agentic AI evolves AI from a tool that supports human work into a delegated organizational actor capable of goal-directed reasoning, adaptive self-regulation, and consequential action-taking. This shift raises a central question: how does agentic AI reshape the dynamics through which organizations accumulate knowledge while preserving belief diversity? Building on and extending March’s (1991) mutual-learning model, we develop an agent-based model that introduces agentic AI as a new class of learner-actor. We examine how its key features affect knowledge accumulation and belief diversity under varying levels of environmental turbulence. Our findings show that agentic AI’s consequential action-taking acts as a turbulence-contingent lever, driving both knowledge and diversity only under unstable conditions. Human reconfiguration of agentic AI sustains diversity in stable environments but becomes redundant under turbulence. In contrast, agentic AI’s self-regulation introduces a knowledge–diversity trade-off that weakens as turbulence increases. These findings contribute to IS and organizational learning research.
Keywords
Agentic AI
organizational learning
knowledge accumulation
belief diversity
agent-based modeling
HSG Classification
contribution to scientific community
Refereed
Yes
Publisher
Association for Information Systems
Event Title
Forty-Seventh International Conference on Information Systems (ICIS)
Event Location
Lisbon, Portugal
Event Date
December 2026
Subject(s)
Division(s)
File(s)![Thumbnail Image]()
open.access
Name
ICIS-2026_-Revision_1974_final.pdf
Size
1.87 MB
Format
Adobe PDF
Checksum (MD5)
55bcc0413ba1adc67b888be6d19b8759