This thesis explores how artificial intelligence (AI) is transforming UX design, with particular attention to the design of functional architectures, the structural logic that connects user goals to system behaviors. While AI-based tools increasingly automate visual and interface design, their ability to support the reasoning processes underlying UX practice remains uncertain. Through the analysis of current professional workflows and AI-based design tools, the research identifies a gap between the automation of surface-level tasks and that of the deeper methodological layers of design. To address this gap, the author develops and introduces AIFAD (AI for Functional Architecture Design), a framework designed to evaluate AI tools for UX from a methodological perspective. The framework defines four key metrics (Logical Consistency, Adherence to Requirements, Ideation Support, and Repetitiveness/Context Adaptability), and five prompt typologies, aiming to investigate in which aspects these tools can be genuinely useful for designers and how linguistic and structural formulation influence their performance. Empirical tests conducted with UX experts on four AI-driven flow-generation tools (Whimsical, Eraser.io, Diagram Generator, and Mermaid Chart) show that these systems can accelerate the early visualization of user flows but still struggle to demonstrate real contextual understanding, solid logical coherence, and creative depth. The results highlight that AI has strong potential to redefine the designer’s role in the ideation phase, acting as a reasoning partner rather than a substitute. However, current tools remain far from achieving this potential, offering schematic outputs and limited interpretive support. The study concludes that collaboration between AI and designers could enrich the ideation phase if guided critically, transforming automation into a means to amplify human intentionality and creativity.
Questa tesi esplora come l’intelligenza artificiale (AI) stia trasformando il design UX, con particolare attenzione alla progettazione delle architetture funzionali, la logica strutturale che connette gli obiettivi degli utenti ai comportamenti del sistema. Mentre gli strumenti basati su AI automatizzano sempre più spesso la progettazione visiva e d’interfaccia, la loro capacità di supportare i processi di ragionamento che stanno alla base della pratica UX rimane incerta. Attraverso l’analisi dei flussi di lavoro in ambienti professionali attuali e degli strumenti di progettazione AI-based, la ricerca individua un divario tra l’automazione dei livelli più superficiali e quella dei livelli metodologici più profondi del design. Per rispondere a questa lacuna, l'autrice sviluppa e introduce AIFAD (AI for Functional Architecture Design), un framework pensato per valutare gli strumenti AI destinati alla UX da una prospettiva metodologica. Il framework definisce quattro metriche principali (Coerenza logica, Aderenza ai requisiti, Supporto all’ideazione e Ripetitività/Adattabilità al contesto) e cinque tipologie di prompt, con l’obiettivo di indagare in quali aspetti questi strumenti possano risultare realmente utili per i designer e in che modo la formulazione linguistica e strutturale influenzi le loro prestazioni. I test empirici, condotti con esperti di UX su quattro strumenti di generazione di flussi AI-driven (Whimsical, Eraser.io, Diagram Generator e Mermaid Chart), mostrano che tali sistemi possono accelerare la visualizzazione preliminare dei flussi utente, ma faticano ancora a dimostrare una reale comprensione del contesto, una coerenza logica solida e una profondità creativa. I risultati evidenziano come l’AI possa avere un forte potenziale nel ridefinire il ruolo del designer nella fase di ideazione, agendo come un partner di ragionamento piuttosto che come un sostituto. Tuttavia, gli strumenti attuali restano lontani dal raggiungere questo potenziale, offrendo output schematici e un supporto interpretativo limitato. Lo studio conclude che la collaborazione tra AI e designer potrebbe arricchire la fase di ideazione se guidata in modo critico, trasformando l’automazione in un mezzo per amplificare l’intenzionalità e la creatività umana.
What changes when functional architecture is designed by AI? A critical study of methodological transformations in UX design and evaluation framework of current tools
De CEGLIA, CHIARA MARIA
2024/2025
Abstract
This thesis explores how artificial intelligence (AI) is transforming UX design, with particular attention to the design of functional architectures, the structural logic that connects user goals to system behaviors. While AI-based tools increasingly automate visual and interface design, their ability to support the reasoning processes underlying UX practice remains uncertain. Through the analysis of current professional workflows and AI-based design tools, the research identifies a gap between the automation of surface-level tasks and that of the deeper methodological layers of design. To address this gap, the author develops and introduces AIFAD (AI for Functional Architecture Design), a framework designed to evaluate AI tools for UX from a methodological perspective. The framework defines four key metrics (Logical Consistency, Adherence to Requirements, Ideation Support, and Repetitiveness/Context Adaptability), and five prompt typologies, aiming to investigate in which aspects these tools can be genuinely useful for designers and how linguistic and structural formulation influence their performance. Empirical tests conducted with UX experts on four AI-driven flow-generation tools (Whimsical, Eraser.io, Diagram Generator, and Mermaid Chart) show that these systems can accelerate the early visualization of user flows but still struggle to demonstrate real contextual understanding, solid logical coherence, and creative depth. The results highlight that AI has strong potential to redefine the designer’s role in the ideation phase, acting as a reasoning partner rather than a substitute. However, current tools remain far from achieving this potential, offering schematic outputs and limited interpretive support. The study concludes that collaboration between AI and designers could enrich the ideation phase if guided critically, transforming automation into a means to amplify human intentionality and creativity.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_12_DeCeglia.pdf
accessibile in internet solo dagli utenti autorizzati
Dimensione
5.64 MB
Formato
Adobe PDF
|
5.64 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/10589/246731