Supply chain resilience has become increasingly important for SMEs exposed to disruptions, yet limited empirical evidence explains how these firms perceive and use Industry 4.0 technologies in practice. This paper investigates how Industry 4.0 technologies are associated with different supply chain resilience dimensions in wine-sector SMEs. Based on nine semi-structured interviews with organizations in Portugal and Italy, the study analyses the role of digital tools in supporting visibility, collaboration, redundancy, flexibility, adaptability, and risk awareness. The findings show that resilience is not primarily driven by advanced or fully integrated Industry 4.0 systems. Instead, SMEs rely on a hybrid model combining relational trust, tacit knowledge, practical routines, lightweight Industry 4.0-related applications, and basic digital supports. Tools such as ERP modules, freight portals, e-commerce and supplier portals supported connectivity, traceability, coordination, and information availability, while spreadsheets, WhatsApp, and direct communication worked mainly as complementary operational supports. The paper contributes to a more context-sensitive understanding of technology-enabled resilience in resource-constrained SME supply chains.
Energy forecasting plays a critical role in ensuring efficiency and cost control across food supply chains, particularly in energy-intensive processes such as industrial freezing and cold storage. However, accurate predictions are often hindered by limited and heterogeneous historical data. This paper presents a benchmarking study of pretrained time series transformer models under few-shot and zero-shot learning settings, applied to energy consumption forecasting in a real-world agrifood pilot. The study evaluates forecasting granularity, data availability, and computational performance over a 117-month industrial dataset. Results show that transformer-based approaches, especially Chronos-based models, can provide competitive forecasts under data scarcity, supporting scalable forecasting workflows for collaborative energy planning between food processing companies and energy providers.
Artificial Intelligence (AI) is being integrated into Production Management (PM) routines. Still, existing research provides fragmented explanations of how Human-AI Collaboration should be governed when operational decisions are distributed amongst human and AI actors. This challenge primarily affects industrial SMEs, where governance structures are informal, and oversight, adaptation, and lifecycle management resources are limited. This paper argues the central problem is not AI adoption per se, but the governance of PM routines in which humans and AI collectively influence operational decisions. Integrating Sociotechnical Systems, Agency, and Coordination theory, the paper conceptualises PM routines as localised Hybrid Collaborative Networks and develops a governance design framework for industrial SMEs. The framework supports the analysis and design of decision rights, escalation, accountability, and adaptation along different PM routine types and AI autonomy levels. It contributes a governance-oriented artefact for anticipatory analysis and future design of adaptive Human-AI Collaboration in industrial SMEs’ PM routines.
As the integration of Artificial Intelligence (AI) increases in the applications we use in our daily work and life, the need for supporting appropriate design of such applications has increased. Recent trends show a shift in collaborating with and trusting AI systems to collaborating with a human-AI symbiotic system. Numerous authors propose design principles and guidelines for such applications, yet there is a lack of specifics and nuances to guide solution designers to ensure a smooth and seamless collaboration between humans and AI systems. This paper proposes a set of principles for designing solutions that promote human-AI symbiosis, where humans and AI systems can work and learn together. The design principles were identified through a review of recent literature, with the aim to provide a practical approach to solution designers. Examples of how these principles could be applied are described. This work has been conducted within the context of two research projects.
Business-technology alignment alongside generative AI (GenAI) represents a paradigm shift with the potential to accelerate digitalization. The question is how the approach can help move from one-of-a-kind, siloed applications integrated via mediation to open, multivendor, adaptive integration development and operations frameworks. This paper discusses a model-based integration, prompting the Application and the Application Programming Interface (API) as central entity-type concepts. The approach relies on Business Service for conceptual modeling and on Computing Service (CService) for implementation modeling towards vendor-lock-in-free applications and technology landscapes. The CService entity logically encapsulates software and resources within an informatics system implementing zero or more Applications or APIs. The positioned approach is to leverage collaborative user-GenAI symbiotic development and operations of both conceptual sociotechnical business systems models and digital technology informatics systems towards open applications and API frameworks. The proposed framework is further discussed in terms of the preparedness of organizations joining Collaborative Networks (CNs).
