Empowering Viva Assessments

We formalised the project scope and objectives, confirmed collaboration with the AI development team, and identified two core user groups: Student Practice, and Question Generation. Initial feasibility discussions around AI-supported viva assessment laid the groundwork for everything that followed.

A multidisciplinary student development team was recruited and onboarded, with formal roles allocated across front-end, back-end, AI integration, UX research, and documentation. We selected our AI model approach, defined the integration strategy, and established the foundational front-end architecture and data flows.

We conducted in-depth interviews with academic staff to surface key concerns and refine our feature set. AI prompt structures were designed with care, and a formal Human Research Ethics Application (HREA) was submitted to ensure our research met institutional standards.

Front-end development commenced alongside the knowledge input ingestion pipeline. We integrated Azure-based secure data storage hosted in Australia and began early structured testing of our AI models — validating core prompt logic ahead of the next build phase.

Regeneration logic was implemented with a hybrid notes system supporting both global and per-item instructions. Transcript handling and an AI-assisted review interface were integrated, the Help & Guides knowledge base was developed for students and staff, and internal stress testing covered file handling, permissions, and AI edge cases.

The team conducted comprehensive internal alpha testing across file input robustness, AI question personalisation, summative workflow integrity, accessibility features, and role-based permissions. Edge cases were identified and documented, the UI was refined based on usability findings, and improved error handling and loading states were shipped.