Speakers
Description
Conventional computerized maintenance management systems (CMMSs) support the documentation and administration of maintenance activities, but they largely depend on users to interpret information, initiate workflows, coordinate resources, and make decisions. This paper proposes an agentic maintenance management system that extends traditional CMMS capabilities through autonomous and goal-directed artificial-intelligence agents. The technical laboratories at the University of Stavanger (UiS) are used as a case study to design and evaluate the proposed system. The study covers four core modules: scheduled maintenance, corrective maintenance, spare-parts management, and notifications. Within these modules, specialized agents monitor maintenance requirements, identify overdue or emerging tasks, prioritize work orders, recommend schedules, check spare-parts availability, support replenishment decisions, and deliver context-sensitive notifications to relevant personnel. A conventional CMMS workflow is modelled as the baseline and compared with the proposed agentic system using process performance, decision quality, responsiveness, traceability, and human-effort indicators. The comparison examines whether agentic capabilities can reduce manual coordination, shorten response and processing times, improve maintenance-plan compliance, increase spare-parts availability, and support timely communication. The proposed architecture retains human oversight for safety-critical, costly, or uncertain decisions, while allowing routine administrative and analytical tasks to be performed autonomously. The case demonstrates how maintenance management can progress from passive data recording and user-triggered transactions toward proactive, adaptive, and decision-oriented support. The paper provides a practical framework for integrating agentic AI with existing maintenance processes and identifies the governance, data-quality, cybersecurity, transparency, and human-control requirements necessary for responsible implementation in technical laboratory environments.
| Topic Area | Smart Operations and Maintenance |
|---|