The growing adoption of artificial intelligence (AI) in supply chain management has emphasized the need for structured frameworks to effectively support AI project implementation. While the literature highlights significant potential benefits, many AI initiatives fail due to organizational and managerial challenges rather than technological limitations. In this context, this thesis aims to empirically validate and, where necessary, refine a theoretical framework designed to guide the implementation of AI projects in supply chain environments. The framework analysed is derived from the study “What could go wrong? Unlocking AI Potential in Supply Chain Management: Design Principles and Best Practices” and is based on Critical Success Factors (CSFs) identified in the literature and organized into principles and conceptual pillars. To assess its robustness and applicability, a qualitative research approach was adopted, based on seven case studies collected through semi-structured interviews with professionals directly involved in AI projects across the supply chain. The empirical analysis followed a two-step approach, combining within-case and cross-case analyses to identify recurring patterns and divergences. The results show strong empirical alignment with the framework, confirming the relevance of several principles as key enablers of AI project success. Principles that were inconsistently applied were often associated with inefficiencies or implementation challenges, further supporting the framework’s validity. In addition, two main deviations from the original framework emerged. First, the “Explainable and Transparent AI” pillar was not significantly adopted in the analysed cases, leading to its reinterpretation as a facilitating rather than enabling element. Second, the findings suggest revising the project timeline by positioning the creation of interdisciplinary teams prior to the resource assessment phase. Overall, this thesis provides an empirically grounded contribution to the understanding of AI implementation in supply chain management, offering both a refinement of existing theoretical frameworks and concrete practical insights. From an academic perspective, the study strengthens the connection between theory and empirical evidence, contributing to the ongoing debate on critical success factors for AI projects. From a managerial standpoint, the findings provide actionable guidance for practitioners, supporting managers and decision-makers in structuring, managing, and successfully implementing AI initiatives within complex supply chain contexts.
La crescente adozione dell’intelligenza artificiale (AI) nella gestione della supply chain ha evidenziato la necessità di framework strutturati in grado di supportare efficacemente l’implementazione dei progetti di AI. Sebbene la letteratura metta in luce i potenziali benefici significativi di tali tecnologie, molte iniziative di AI falliscono a causa di criticità di natura organizzativa e manageriale piuttosto che per limiti tecnologici. In questo contesto, la presente tesi si pone l’obiettivo di validare empiricamente e, ove necessario, affinare un framework teorico progettato per guidare l’implementazione di progetti di intelligenza artificiale in ambito supply chain. Il framework analizzato deriva dallo studio “What could go wrong? Unlocking AI Potential in Supply Chain Management: Design Principles and Best Practices” ed è basato sui Critical Success Factors (CSF) identificati nella letteratura e organizzati in principi e pilastri concettuali. Per valutarne la solidità e l’applicabilità, è stato adottato un approccio di ricerca qualitativo, fondato sull’analisi di sette case study raccolti tramite interviste semi-strutturate a professionisti direttamente coinvolti in progetti di AI lungo la supply chain. L’analisi empirica ha seguito un approccio in due fasi, combinando within-case analysis e cross-case analysis, al fine di individuare pattern ricorrenti e differenze significative tra i casi. I risultati mostrano una forte aderenza empirica al framework, confermando la rilevanza di diversi principi come fattori abilitanti chiave per il successo dei progetti di AI. I principi applicati in modo non coerente risultano spesso associati a inefficienze o difficoltà di implementazione, fornendo ulteriore supporto alla validità del framework. Sono inoltre emerse due principali deviazioni rispetto alla formulazione originale del framework. In primo luogo, il pilastro “Explainable and Transparent AI” non è stato adottato in modo significativo nei casi analizzati, portando a una sua reinterpretazione come elemento facilitatore piuttosto che abilitante. In secondo luogo, i risultati suggeriscono una revisione della timeline progettuale, collocando la creazione di team interdisciplinari in una fase antecedente alla valutazione delle risorse. Nel complesso, questa tesi fornisce un contributo empiricamente fondato alla comprensione dell’implementazione dell’AI nella supply chain, offrendo sia un affinamento dei framework teorici esistenti sia indicazioni pratiche concrete. Dal punto di vista accademico, lo studio rafforza il collegamento tra teoria ed evidenza empirica, contribuendo al dibattito sui fattori critici di successo dei progetti di AI. Dal punto di vista manageriale, i risultati forniscono linee guida operative utili a manager e decision-maker per strutturare, gestire e implementare con successo iniziative di intelligenza artificiale in contesti complessi di supply chain.
Implementing Artificial Intelligence system in Supply Chain management - AI design principlesframework application
Varè, Luca
2024/2025
Abstract
The growing adoption of artificial intelligence (AI) in supply chain management has emphasized the need for structured frameworks to effectively support AI project implementation. While the literature highlights significant potential benefits, many AI initiatives fail due to organizational and managerial challenges rather than technological limitations. In this context, this thesis aims to empirically validate and, where necessary, refine a theoretical framework designed to guide the implementation of AI projects in supply chain environments. The framework analysed is derived from the study “What could go wrong? Unlocking AI Potential in Supply Chain Management: Design Principles and Best Practices” and is based on Critical Success Factors (CSFs) identified in the literature and organized into principles and conceptual pillars. To assess its robustness and applicability, a qualitative research approach was adopted, based on seven case studies collected through semi-structured interviews with professionals directly involved in AI projects across the supply chain. The empirical analysis followed a two-step approach, combining within-case and cross-case analyses to identify recurring patterns and divergences. The results show strong empirical alignment with the framework, confirming the relevance of several principles as key enablers of AI project success. Principles that were inconsistently applied were often associated with inefficiencies or implementation challenges, further supporting the framework’s validity. In addition, two main deviations from the original framework emerged. First, the “Explainable and Transparent AI” pillar was not significantly adopted in the analysed cases, leading to its reinterpretation as a facilitating rather than enabling element. Second, the findings suggest revising the project timeline by positioning the creation of interdisciplinary teams prior to the resource assessment phase. Overall, this thesis provides an empirically grounded contribution to the understanding of AI implementation in supply chain management, offering both a refinement of existing theoretical frameworks and concrete practical insights. From an academic perspective, the study strengthens the connection between theory and empirical evidence, contributing to the ongoing debate on critical success factors for AI projects. From a managerial standpoint, the findings provide actionable guidance for practitioners, supporting managers and decision-makers in structuring, managing, and successfully implementing AI initiatives within complex supply chain contexts.| File | Dimensione | Formato | |
|---|---|---|---|
|
Implementing Artificial Intelligence systems in Supply Chain Management - AI design principles framework application.pdf
accessibile in internet per tutti
Dimensione
3.06 MB
Formato
Adobe PDF
|
3.06 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/250871