Digital Transformation in SMEs: A Data-driven Framework for Explainable AI-Enabled Decision Support in Supply Chain Resilience

The contemporary business environment is characterized by unprecedented volatility, uncertainty, complexity, and ambiguity (VUCA), a reality that has exposed the fragility of supply chains across all organizational scales. Small and medium-sized enterprises (SMEs), which constitute the backbone of global economies, are particularly vulnerable to supply chain disruptions due to their limited resources, less formalized risk management practices, and lower technological maturity. Digital transformation offers a strategic pathway for SMEs to enhance supply chain resilience (SCR), yet adoption remains slow because of perceived complexity, cost constraints, and a lack of interpretable, actionable decision support tools. This article addresses these gaps by proposing a novel, data-driven framework that integrates explainable artificial intelligence (XAI) into a decision support system tailored specifically for SME contexts. The framework synthesizes principles from resilience engineering, information systems theory, and SME capability literature to produce a layered architecture comprising data ingestion, predictive analytics, explainable modeling, and prescriptive recommendation modules. Unlike black-box AI systems that generate predictions without justification, our framework leverages SHAP (SHapley Additive ex Planations) values, LIME (Local Interpretable Model-agnostic Explanations), and counterfactual explanations to render model outputs transparent, thereby fostering trust and facilitating informed managerial action. Through a design science research methodology, we develop and validate the framework using both simulated supply chain data and real-world transactional data from a panel of manufacturing SMEs. Empirical evaluation demonstrates that the explainable AI module not only improves forecast accuracy for disruption events by ۲۳% compared to baseline statistical methods but also significantly enhances managerial decision quality, as measured by reduced response times and increased confidence in resource reallocation. The article contributes a theoretically grounded, empirically tested framework that demystifies AI for SME decision makers, operationalizes resilience as a measurable and improvable construct, and provides actionable guidance for technology vendors, policy makers, and SME consortia seeking to accelerate digital transformation in supply chain management.

کلیدواژه ها:

Digital transformation ، small and medium-sized enterprises (SMEs) ، explainable AI (XAI) ، supply chain resilience ، decision support systems ، data-driven framework