Leveraging Generative AI for Knowledge Capture and Management in Manufacturing Small and Medium Enterprises: A Framework for Training, Troubleshooting, and Operational Resilience

Abstract

Manufacturing small and medium enterprises (SMEs) face a critical and accelerating knowledge management crisis. As the baby boomer workforce approaches retirement age, an estimated 70% of critical manufacturing knowledge remains undocumented—embedded in the experience of veteran employees whose departure creates operational vulnerabilities, including quality degradation, extended onboarding timelines, and increased warranty exposure. This article proposes a generative artificial intelligence (GenAI)-powered knowledge management framework specifically designed for manufacturing SMEs, addressing the institutional gap between the knowledge capture needs of lean operations and the capabilities of enterprise-grade knowledge management systems. The proposed framework integrates a retrieval-augmented generation (RAG)-based large language model (LLM) architecture with a structured data collection and human validation pipeline, deploying three functional modules: an interactive training assistant, an AI-powered troubleshooting guide, and a quick reference portal. Drawing on a representative case application in a manufacturing SME experiencing service technician knowledge loss, the framework demonstrates measurable improvements in first-time fix rates, warranty claim reduction, and technician onboarding efficiency. Quantitative evidence from validated literature indicates RAG-based industrial knowledge management systems achieve mean reciprocal rank (MRR) scores of 88–98% and knowledge retrieval recall exceeding 85%. The framework offers a scalable, cost-effective pathway for manufacturing SMEs to preserve institutional knowledge, reduce workforce dependency, and build operational resilience in the face of accelerating workforce transitions.