Agentic AI for Change Management: A Case Study of the AI Change Management Process Assessment for Medical Device Manufacturers

St John Lynch, Niamh and Mc Hugh, Martin and Loughran, Roisin and McCaffery, Fergal (2026) Agentic AI for Change Management: A Case Study of the AI Change Management Process Assessment for Medical Device Manufacturers. Design and Artificial Intelligence. (Submitted)

PDF (Abstract: Agentic AI for Change Management: A Case Study of the AI Change Management Process Assessment for Medical Device Manufacturers) - Accepted Version Download (1MB)

Abstract

Medical device manufacturers are rapidly integrating Artificial Intelligence (AI) into software-as-a-medical-device (SaMD) and AI-enabled devices (AIeMD), which amplifies regulatory, safety and governance obligations. This paper demonstrates how an agentic AI system can autonomously execute the AI Change Management Process (CMP) Assessment. The AI-CMP assessment is an integrated, 35-section change assessment that spans medical device identification, risk classification, software safety classification and incorporates requirements from software lifecycle (IEC 62304), AI governance, data management (IEC PAS 63621), model development and testing (BS 30440 and IEC 63450-draft), product and software risk management (ISO 14971, ISO 24971-2 and PD IEC TR 80002-1), cybersecurity (IEC 81001-5-1), EU AI Act Articles 9–15 (2024/1689), and post-market monitoring (ISO 13485). Using a case study of a new-start manufacturer (MemoryTell), we propose that an agentic AI reduces assessment time by 68%, raises documentation completeness from 62% to 98%, and achieves full traceability from hazards to controls and verification. The system preserves human oversight for clinical and regulatory decisions, aligning with the EU AI Act’s transparency and oversight provisions. The study suggests agentic AI functions as a force-multiplier for quality and regulatory teams, improving safety, consistency and audit readiness while maintaining manufacturer accountability. A mixed-methods and design research approach is taken to develop a State-of-the-Art change assessment tool that can be utilised for AI Agent adoption. Real-world use case demonstrates validity of the AI-CMP tool using qualitative and quantitative data analysis by large and small medical device manufacturers supporting adoption readiness and potential for development of an AI-agent.

Item Type:	Article
Uncontrolled Keywords:	Software, Medical Device, Artificial Intelligence, Agentic AI, AI Agent, Change Management.
Subjects:	Business Engineering Computer Science > Computer Software
Research Centres:	Regulated Software Research Centre
Depositing User:	Niamh StJohnLynch
Date Deposited:	24 Mar 2026 16:44
Last Modified:	24 Mar 2026 16:44
License:	Creative Commons: Attribution-Noncommercial-Share Alike 4.0
URI:	https://eprints.dkit.ie/id/eprint/1034

Actions (login required)

View Item

Downloads