Integrasi Fuzzy Kano Model pada Proses Desain Pengalaman Pengguna Web App dengan Generative AI

Praadita, Firman Noor (2026) Integrasi Fuzzy Kano Model pada Proses Desain Pengalaman Pengguna Web App dengan Generative AI. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032242115-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (11MB) | Request a copy

Abstract

Perkembangan Generative AI (Gen AI) membuka peluang akselerasi proses desain pengalaman pengguna, namun belum banyak penelitian yang mendokumentasikan protokol integrasinya secara end-to-end dalam alur kerja desain, khususnya dengan mekanisme validasi yang memastikan keluaran AI tetap berorientasi pada kebutuhan pengguna. Penelitian ini mengeksplorasi dan mendokumentasikan protokol integrasi desain berbantuan AI dengan fuzzy Kano Model sebagai validation checkpoint untuk memastikan fitur hasil kurasi Gen AI selaras dengan preferensi pengguna aktual. Objek penelitian adalah sistem Simponia, platform showcase portofolio dan manajemen komunitas berbasis web untuk mahasiswa Program Studi Informatika Universitas Muhammadiyah Malang yang memerlukan re-engineering menyeluruh. Penelitian mengadopsi mixed methods dengan kerangka Think-Design-Check dalam satu siklus end-to-end. Fase Think menggunakan Claude Opus melalui pendekatan single-stage prompting untuk menghasilkan 36 fitur kandidat yang dievaluasi berdasarkan tiga dimensi ISO 25010, yaitu Functional Suitability, Operability, dan Maintainability. Validasi fitur dilakukan melalui fuzzy Kano Model berdasarkan masukan 87 responden mahasiswa, menghasilkan 11 fitur prioritas pada Kuadran 1 yang memenuhi kriteria kelayakan teknis AI sekaligus preferensi pengguna. Fase Design mentransformasikan fitur prioritas menjadi prototipe high-fidelity interaktif menggunakan Figma Make melalui global context prompting dalam empat siklus iterasi. Fase Check mengevaluasi usabilitas prototipe menggunakan System Usability Scale (SUS) terhadap 31 responden, menghasilkan skor rata-rata 64,84 yang berada pada kategori Marginal dan belum mencapai threshold ≥ 68. Temuan utama menunjukkan bahwa fuzzy Kano Model efektif mendeteksi divergensi antara penilaian AI dan preferensi pengguna, dengan 65,6% fitur yang dinilai tinggi oleh AI (HIGH PRIORITY dan CONSIDER) ternyata dikategorikan Indifferent oleh pengguna. Penelitian ini menghasilkan protokol yang terdokumentasi dan dapat direplikasi untuk desain pengalaman pengguna berbantuan Gen AI dengan dual validation framework, memberikan kontribusi metodologis pada literatur kolaborasi manusia-AI dalam proses kreatif, serta menyediakan rekomendasi implementasi berbasis bukti empiris untuk pengembangan sistem Simponia.
=======================================================================================================================================
The advancement of Generative AI (Gen AI) presents opportunities to accelerate user experience design processes; however, limited research has documented end-to-end integration protocols within design workflows, particularly with validation mechanisms that ensure AI outputs remain aligned with actual user needs. This study explores and documents an AI-assisted design protocol integrating the Fuzzy Kano Model as a validation checkpoint to ensure Gen AI-curated features align with actual user preferences. The research object is Simponia, a web-based portfolio showcase and community management platform for students of the Informatics Study Program at Universitas Muhammadiyah Malang requiring comprehensive re-engineering. The study adopts a mixed methods approach within a Think-Design-Check framework executed in a single end-to-end cycle. The Think phase employed Claude Opus through a single-stage prompting approach to generate 36 feature candidates evaluated against three ISO 25010 dimensions: Functional Suitability, Operability, and Maintainability. Feature validation was conducted through the Fuzzy Kano Model based on input from 87 student respondents, yielding 11 priority features in Quadrant 1 that satisfied both AI technical feasibility criteria and user preferences simultaneously. The Design phase transformed priority features into an interactive high-fidelity prototype using Figma Make through global context prompting across four iterative cycles. The Check phase evaluated prototype usability using the System Usability Scale (SUS) with 31 respondents, producing a mean score of 64.84 categorized as Marginal and falling short of the ≥68 threshold. Key findings demonstrate that the Fuzzy Kano Model effectively detects divergence between AI assessments and user preferences, with 65.6% of features rated highly by AI (HIGH PRIORITY and CONSIDER) categorized as Indifferent by users. This research produces a documented and replicable protocol for Gen AI-assisted user experience design using a dual validation framework, contributing methodologically to the literature on human-AI collaboration in creative processes, and providing empirically grounded implementation recommendations for Simponia system development.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Generative AI, Desain Pengalaman Pengguna, Fuzzy Kano Model, System Usability Scale, Kolaborasi Manusia-AI
Subjects:	T Technology > T Technology (General) > T59.7 Human-machine systems.
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Firman Noor Praadita
Date Deposited:	27 Jul 2026 04:06
Last Modified:	27 Jul 2026 04:06
URI:	http://repository.its.ac.id/id/eprint/136311

Actions (login required)

View Item