Design and development of a senior-friendly mobile ecommerce app with AI voice support

Lim, Jia Huey (2026) Design and development of a senior-friendly mobile ecommerce app with AI voice support. Final Year Project, UTAR.

PDF Download (12Mb)

Abstract

Although the growth of mobile e-commerce in Malaysia is rapid, the level of improvement in the accessibility of these applications for elderly people remains low. Most popular mobile e-commerce applications in Malaysia, such as Shopee, Lazada, and GrabMart, are developed with the interests of the younger population in mind. As such, these applications pose three challenges to the elderly population aged 60 and above. These challenges include small text and low contrast, which affect the readability of the applications; dependence on the touch interface, which requires high precision; and the high cognitive load required to complete complex steps. As the population in Malaysia is ageing, with Malaysia's elderly population forecast to reach 15.3% of the total population by 2030, the underrepresentation of older adults in mobile commerce has emerged as a pressing societal concern. This project developed a voice-based mobile e-commerce application called SilverCart targeting the elderly population in Malaysia. The project employed the Rapid Application Development (RAD) methodology, which is divided into four phases. These phases include the requirements planning phase, the user design phase, the construction and feedback phase, and the cutover and implementation phase. The application was developed using React Native with Expo SDK 54 and TypeScript. It is powered by Supabase PostgreSQL with Row Level Security. The voice navigation feature was enabled through the incorporation of the Google Gemini API, which is based on the 'gemini-1.5-flash' library. It utilizes function calls to process voice-based natural language inputs. The application also ensured compliance with the AAA standards set by the Web Content Accessibility Guidelines. It achieved this by using a minimum font size of 18pt and touch targets no smaller than 60dp × 60dp. This voice navigation feature was tested functionally to ensure that it could support single-item cart additions, multiple-item additions via one voice command, and trusted contact calls via voice only. It also successfully implemented trust and safety features, including a five-minute post-checkout cancellation window, trusted contact SMS, and an emergency help button. The application was also successfully compiled into a deployable Android APK with 54 screens and three navigation stacks. SilverCart demonstrates that a fully functional, voice-driven e-commerce application can be built to serve elderly users in Malaysia. The project establishes a practical foundation for bridging the accessibility gap that currently exists in the Malaysian mobile commerce ecosystem. It also demonstrates the viability of the application in addressing the current gap in the Malaysian e-commerce landscape.

Actions (login required)