Database-Centered AI Architectures for Smart Manufacturing: A Product Lifecycle Systems Perspective
Main article
Abstract
Smart manufacturing has advanced rapidly, yet the operational value of artificial intelligence is still constrained by fragmented data architectures, uneven interoperability, and weak lifecycle integration. This article reframes smart manufacturing from a database-centered perspective and argues that the decisive issue is no longer whether firms possess AI tools, but whether they can organize lifecycle data into architectures that support coordination, traceability, governance, and learning. Building on a Product Lifecycle Management (PLM) perspective, the study develops a structured conceptual architecture that links four lifecycle domains—strategy and organization, value-chain intelligence, management support, and infrastructure and capabilities—to five data-architecture requirements: interoperability, traceability, governance, decision coupling, and scalability. To make the argument operational, the article conducts a secondary analytical synthesis of the smart-manufacturing literature and translates the major themes into a comparative coding matrix. The results show that infrastructure capabilities and strategy layers require the strongest attention to interoperable identifiers, governance rules, and scalable event pipelines, while value-chain and management processes depend more heavily on closed-loop decision coupling and cross-functional traceability. A lifecycle-aligned architecture is then proposed in which data fabric, semantic modeling, AI services, and feedback control are coordinated as a coherent computational system rather than implemented as isolated modules. The article contributes by moving the smart-manufacturing discussion from technology inventories toward database-aware systems design, by offering a practical coding framework for evaluating lifecycle data readiness, and by identifying an enterprise roadmap for AI-enabled manufacturing transformation. The central implication is that smart manufacturing should be designed as a database-centered system of lifecycle intelligence in which AI models, operational workflows, and governance mechanisms are co-developed.
