EUVIMEDEUVIMED entdecken
Crossref

Integrating Generative AI, Prompt-to-Result Interaction Patterns, and Adaptive UX for AI-Native Mobile Applications: A Design Engineering Framework

Gazi Yeasin Arifin

International Journal of Scientific Interdisciplinary Research · 01.12.2025 · 06 (02) · 173-202

DOI: 10.63125/n5z7gv27 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Scientific Interdisciplinary Research, online erschienen am 01.12.2025. Verfasst von Gazi Yeasin Arifin.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

AI-native mobile applications increasingly depend on generative artificial intelligence, yet inconsistent output quality, weak prompt guidance, limited adaptation, and opaque system behavior can undermine user interaction. This study examined how five interface characteristics relate to perceived user experience effectiveness and developed a Task-Technology Fit-informed design engineering framework for prompt-to-result workflows. A quantitative, cross-sectional, comparative four-case design was embedded within a convergent mixed-methods approach incorporating five-point Likert questionnaires, task walkthroughs, and interviews. The manuscript reports 320 valid questionnaire responses, with 80 respondents in each of four AI-native mobile application cases, alongside 48 walkthrough participants and 24 interviewees. The independent variables comprised Generative AI Output Quality, Prompt Guidance and Interaction Affordances, Adaptive UX Responsiveness, Feedback and Iterative Revision Control, and AI Interaction Transparency and Predictability; the outcome was Perceived User Experience Effectiveness. Descriptive statistics, Cronbach’s alpha, measurement diagnostics, Pearson correlation, and hierarchical multiple regression were combined with thematic and cross-case analysis. Reported reliability coefficients ranged from 0.839 to 0.869, with KMO = 0.900. Perceived effectiveness averaged 3.81 on a five-point scale. All five predictors were positively correlated with effectiveness (r = 0.443–0.546, p < .001). The full regression model explained 56.4% of variance (adjusted R² = 0.550), with prompt guidance showing the strongest standardized association (β = 0.297), followed by output quality (β = 0.267), adaptive responsiveness (β = 0.245), revision control (β = 0.151), and transparency (β = 0.143). Across the task assessments, 38 of 48 participants completed assigned tasks successfully (79.2%). These reported patterns informed a framework prioritizing guided prompting, relevant outputs, adjustable adaptation, accessible revision, and transparent feedback. The findings suggest priorities for developers and UX engineers, while the observational design limits causal inference and the reported recruitment figures require confirmation against original field records. Future experimental testing should assess task accuracy, interaction efficiency, usability, and user agency.

Zitieren

APA

Arifin, G. Y. (2025). Integrating Generative AI, Prompt-to-Result Interaction Patterns, and Adaptive UX for AI-Native Mobile Applications: A Design Engineering Framework. International Journal of Scientific Interdisciplinary Research, 06(02), 173-202. https://doi.org/10.63125/n5z7gv27

Vancouver

Arifin GY. Integrating Generative AI, Prompt-to-Result Interaction Patterns, and Adaptive UX for AI-Native Mobile Applications: A Design Engineering Framework. International Journal of Scientific Interdisciplinary Research. 2025;06(02):173-202. doi:10.63125/n5z7gv27

BibTeX
@article{Arifin2025,
  author = {Arifin, Gazi Yeasin},
  title = {Integrating Generative AI, Prompt-to-Result Interaction Patterns, and Adaptive UX for AI-Native Mobile Applications: A Design Engineering Framework},
  journal = {International Journal of Scientific Interdisciplinary Research},
  year = {2025},
  volume = {06},
  number = {02},
  pages = {173--202},
  doi = {10.63125/n5z7gv27}
}
Kostenlos starten →