The work referred to throughout this section. Hover any card for the full backgrounder.
How my five expertise areas developed across my bachelor, internship and Final Bachelor Project.
Viewing Year 1.
Technology and Realisation was not initially my strongest expertise area, but I developed significantly during Project 3. I was responsible for the electronics setup of our prototype, using an M5Stack Atom microcontroller and a UHF RFID scanner. This allowed the system to respond to children’s actions without invasive monitoring. The process taught me the importance of testing early, researching compatibility and understanding how technical choices affect the final user experience.
UHF RFID scanner and M5 Stack components from Project 3
During my internship, my relationship with technology shifted from building systems myself to understanding how technical decisions scale within a large organisation. Working with engineers and quality teams showed me that technology must be robust, maintainable and integrated into a wider product ecosystem. I began to see technology less as a feature to showcase, and more as an enabling layer for real user value.
Technology and Realisation developed strongly during HoldOn because one of my FBP goals was to prototype and test the most promising direction in a way that allowed realistic evaluation. I achieved this by moving beyond a visual concept and building a working AI-mediated call-screening prototype using Twilio. People could phone a real number, speak to an AI receptionist and experience how the system might intercept and assess a call before it reached the older adult.
System Diagram for HoldOn
The first prototype created a minimum viable technical loop: it intercepted a call, asked the caller for their name and reason for calling, transcribed the response, and used AI to classify the risk level. Later iterations explored behaviours for low-, medium- and high-risk calls, trusted contacts, personalisation and caller verification.
This goal did not go exactly as planned. I initially imagined two neat rounds of testing with at least five users per round, but recruiting older adults was slower and more complex than expected. However, the project still involved older adult interviews, expert feedback, a bubble group and caller-side evaluation with people interacting directly with the AI receptionist. This gave the prototype broader validation than I first expected.
The most important learning was that technology should be used selectively. The value of HoldOn was not that it used AI, but that the technology created a protective pause before scam pressure reached the user and could easily integrate within the user’s existing setup and lifestyle. This strengthened my ability to connect technical feasibility to social and experiential purpose.
