Elevating Mobile User Experience through Artificial Intelligence

Abstract

This work presents a proof-of-concept system demonstrating the technical feasibility of using large language models (LLMs) to generate adaptive mobile user interfaces. The project implements a complete pipeline: a Flutter client that renders interfaces from JSON contracts, a NestJS backend managing user events and contract generation via Google Gemini 2.5 Pro, and a React dashboard for monitoring and control. The system successfully demonstrates end-to-end functionality through two simulated test scenarios. While this PoC establishes architectural viability and technical correctness, empirical validation of user experience improvements remains future work requiring controlled user studies with real participants.