The transition from Industry 4.0 to Industry 5.0 presents manufacturing companies with a fundamental paradox: despite the massive data availability enabled by the Industry 4.0 revolution, many organizations still struggle to translate information abundance into real operational value supporting decision-making. Industry 5.0 intensifies this challenge by emphasizing the integration of digital intelligence into human-centric processes that enhance efficiency and sustainability. While prior research has examined Industry 4.0 technologies and emerging XAI applications in manufacturing, the literature lacks an integrated approach addressing the interconnected challenges preventing data valorization in this context. Indeed, barriers such as Artificial Intelligence (AI) opacity limiting trust, partial evidence of tangible benefits, fragmented technology adoption perspectives, and absence of economic evaluation models are numerous and deeply interconnected. These interdependencies create a reinforcing cycle where one limitation amplifies the others, preventing organizations from fully realizing the benefits of digital transformation. This thesis addresses these challenges by examining the process of transforming data into actionable insights in manufacturing operations contexts across its key stages, offering a structured response to what is defined as the Data Valorization Paradox - the self-reinforcing cycle of barriers that prevent organizations from turning data abundance into tangible value. This research, first, maps how data is currently used within manufacturing operations, along with its current interpretability implications, to assess its effective use in decision-making. Then, it explores how the adoption of cutting-edge enabling technologies can strengthen both operational and sustainable performance, enabling more informed and strategic adoption. Next, experimental applications across semiconductor, predictive maintenance (PdM), and injection molding contexts validate that explanation-generating mechanisms transform opaque AI predictions into actionable operational insights. Finally, the research introduces an economic evaluation framework that identifies the conditions under which investments in these transparency-enhancing mechanisms become not only technologically feasible but also economically justified within manufacturing environments. This Ph.D. thesis contributes to theory by introducing the Data Valorization Paradox as a novel construct for understanding manufacturing digitalization challenges and provides practitioners with actionable guidance on technology prioritization and investment decision-making.
Il passaggio dall’Industry 4.0 all’Industry 5.0 presenta alle aziende manifatturiere un paradosso fondamentale: nonostante la massiccia disponibilità di dati resa possibile dalla rivoluzione Industry 4.0, molte imprese continuano a faticare nel trasformare l’abbondanza di informazioni in reale valore operativo a supporto dei processi decisionali. L’ Industry 5.0 intensifica questa sfida, enfatizzando l'integrazione dell'intelligenza digitale nei processi decisionali centrati sulle persone, a miglioramento dell’efficienza e la sostenibilità. Sebbene la ricerca precedente abbia esaminato le tecnologie dell'Industry 4.0 e le emergenti applicazioni dell'XAI nel settore manifatturiero, in letteratura manca un approccio integrato che affronti le sfide interconnesse che impediscono la valorizzazione del dato in questo contesto. Infatti, barriere quali opacità dell’Intelligenza Artificiale (IA) che limita la fiducia, l’evidenza parziale di benefici tangibili, prospettive frammentate sull’adozione tecnologica e l’assenza di modelli di valutazione economica sono numerose e profondamente interconnesse. Tali interdipendenze generano un circolo auto-rinforzante in cui ogni limitazione amplifica le altre, impedendo alle organizzazioni di cogliere pienamente i benefici della trasformazione digitale. Questa tesi affronta tali sfide esaminando il processo di trasformazione del dato in intuizioni operative e concrete nei contesti manifatturieri, attraverso le sue fasi principali, offrendo una risposta strutturata a quello definito come il Paradosso della Valorizzazione dei Dati - un circolo auto-rinforzante di barriere che impedisce alle aziende di convertire l’abbondanza di dati in valore tangibile. La ricerca fornisce inizialmente una mappatura rispetto all’uso corrente del dato nei contesti manufatturieri e le sue implicazioni di interpretabilità, al fine di valutarne l’uso nel processo decisionale. Successivamente, esplora come l’adozione di tecnologie abilitanti possa rafforzare le prestazioni operative e sostenibili, consentendo un’adozione più informata e strategica. In seguito, applicazioni sperimentali nei contesti di semiconduttori, manutenzione predittiva e stampaggio a iniezione dimostrano che i meccanismi generativi di spiegazioni trasformano le predizioni opache generate dall’IA in intuizioni operative utilizzabili. Infine, viene introdotto un framework di valutazione economica che identifica le condizioni in cui gli investimenti in questi meccanismi di trasparenza risultano non solo tecnologicamente realizzabili, ma anche giustificati dal punto di vista economico nei contesti manifatturieri. Questa tesi contribuisce alla teoria introducendo il Paradosso della Valorizzazione dei Dati come costrutto innovativo per comprendere le sfide della digitalizzazione del settore manifatturiero, fornendo ai professionisti una guida concreta sulla prioritizzazione tecnologica e sulle decisioni d’investimento.
From data generation to value creation: unraveling the data valorization paradox in manufacturing operations
PRESCIUTTINI, ANNA
2025/2026
Abstract
The transition from Industry 4.0 to Industry 5.0 presents manufacturing companies with a fundamental paradox: despite the massive data availability enabled by the Industry 4.0 revolution, many organizations still struggle to translate information abundance into real operational value supporting decision-making. Industry 5.0 intensifies this challenge by emphasizing the integration of digital intelligence into human-centric processes that enhance efficiency and sustainability. While prior research has examined Industry 4.0 technologies and emerging XAI applications in manufacturing, the literature lacks an integrated approach addressing the interconnected challenges preventing data valorization in this context. Indeed, barriers such as Artificial Intelligence (AI) opacity limiting trust, partial evidence of tangible benefits, fragmented technology adoption perspectives, and absence of economic evaluation models are numerous and deeply interconnected. These interdependencies create a reinforcing cycle where one limitation amplifies the others, preventing organizations from fully realizing the benefits of digital transformation. This thesis addresses these challenges by examining the process of transforming data into actionable insights in manufacturing operations contexts across its key stages, offering a structured response to what is defined as the Data Valorization Paradox - the self-reinforcing cycle of barriers that prevent organizations from turning data abundance into tangible value. This research, first, maps how data is currently used within manufacturing operations, along with its current interpretability implications, to assess its effective use in decision-making. Then, it explores how the adoption of cutting-edge enabling technologies can strengthen both operational and sustainable performance, enabling more informed and strategic adoption. Next, experimental applications across semiconductor, predictive maintenance (PdM), and injection molding contexts validate that explanation-generating mechanisms transform opaque AI predictions into actionable operational insights. Finally, the research introduces an economic evaluation framework that identifies the conditions under which investments in these transparency-enhancing mechanisms become not only technologically feasible but also economically justified within manufacturing environments. This Ph.D. thesis contributes to theory by introducing the Data Valorization Paradox as a novel construct for understanding manufacturing digitalization challenges and provides practitioners with actionable guidance on technology prioritization and investment decision-making.| File | Dimensione | Formato | |
|---|---|---|---|
|
Final_Thesis_Anna_Presciuttini.pdf
accessibile in internet per tutti
Dimensione
14.8 MB
Formato
Adobe PDF
|
14.8 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/10589/255157