Featured work
All work →
Testing a More Flexible KYC Flow with Autosave
Tested two KYC approaches to determine how Autosave should work. A randomized usability test with 30 users found no significant difference in SUS or misclicks, while Flow B showed stronger descriptive and qualitative signals and was selected as the implementation baseline.
QRIS (Quick Response Code Indonesian Standard) Top Up
Designed a QRIS Top Up flow using a merchant-presented pattern, allowing users to generate a QR for someone else to scan instead of sharing or entering account details. It later became one of AstraPay's largest cash-in channels.
Testing KYC Hook Copy in Production
Tested three KYC messaging approaches in production to increase funnel entry. All three variants significantly outperformed the legacy copy, with Monetary Highlight selected for its stronger contribution to KYC volume.
Pedagogical Agent Visibility × Task Type: Effects on Cognitive Load & Learning Outcomes
Tested how pedagogical agent visibility affects learning outcomes across psychomotor and cognitive tasks through a 2×2 mixed factorial study with 157 participants. Visible agents improved performance and self-efficacy in psychomotor tasks, while effects on cognitive tasks were more mixed.
