Featured case study Live website ↗ iOS App ↗
0-to-1 digital infrastructure
I founded ASA after noticing a specific gap: Australia already has mental health and community resources for multicultural populations, but for many people from Asian backgrounds, language, cultural context, and how those services are structured still create real distance from the support that exists. What started as a community response quickly became a product problem: the organisation needed more than a public website. It needed a bilingual, trust-building platform that could support content, events, membership, courses, and internal operations. I defined the product structure, system boundaries, and delivery priorities, then used AI as a practical execution partner to build the organisation's digital infrastructure from 0 to 1.
Priority 1
I started by mapping who needed what from the platform: public users, volunteers, lecturers, internal operators, and managers. That made it easier to separate trust and participation paths on the front-end from maintenance and operations on the back-end.
Priority 2
The core user group are overseas Chinese users who need both emotional safety and practical clarity. I treated Chinese and English switching as a day-one product requirement — not a toggle to add later. This decision also directly drove WCAG 2.2 AA compliance: correct html lang declaration, accessible language toggle labelling, and consistent semantic structure across both languages.
Priority 3
The platform had to help the team run the organisation, not only show it. I defined backend modules around real operating objects such as members, publications, course records, certificates, and event publishing so the system could replace fragmented manual work.
Priority 4
I treated the course system as both a current association tool and a potential SaaS direction for future teacher collaboration, while evaluating the app as a lighter-touch member and participation surface. That let me keep the first release focused without losing the longer-term roadmap.
I framed the work as an organisational digital infrastructure initiative rather than a simple website request.
I mapped stakeholders, defined public and operational layers, and built the information architecture around user questions and service logic.
I set the first-release priorities around bilingual trust, core participation paths, backend maintenance capacity, and a stable system skeleton for future growth.
I used AI tools to improve implementation speed across product design, coding, debugging, and structured iteration.
I checked whether implementation still matched the original product intent, then drove the work through to a live organisation platform.
A trustworthy platform needs a coherent visual identity that a distributed volunteer organisation can apply consistently — without a full-time design team. ASA's brand system was designed by a designer board member, with my involvement in setting requirements, giving direction, and making decisions on how the brand would work across the platform.
The journal work was different: I served as Editor-in-Chief of the ASA Mental Health Review and was directly involved in the design and editorial output. That involved structuring the operation from scratch — recruiting a managing editor, section editors, layout designers, and a four-person advisory board covering healing, medicine, psychology, and publishing.
This project also exposed where my first assumptions were too narrow. I initially thought in terms of building a better website, but the real challenge was product structure: account logic, membership levels, course access, operational workflows, and cross-surface data had to be designed intentionally much earlier.
I also learned that “AI-assisted delivery” does not reduce product responsibility. It actually raises the bar on judgment. I had to keep checking whether implementation still matched the platform logic I intended, especially when data flow, hosting, and system boundaries were not fully visible from the interface level.
The result was a deeper shift in how I work: I now treat platform boundary definition, infrastructure risk, and execution verification as part of product design, not as something to think about only after launch.
One concrete post-launch change came from volunteer feedback: the standards and verification pages were not surfaced early enough in the navigation flow. Users arriving from event promotions had no quick way to assess the organisation's legitimacy before deciding to join. I moved the standards and verification entry point higher in the site structure and added clearer trust signals to the homepage — a navigation fix driven by real user behaviour, not by aesthetic preference.
How this project taught me to verify product logic even when the database and backend implementation were not fully visible.
Why Soulful Asia had to move beyond a website and become a structured platform with public and operational layers.
What I learned when the live site went down and I had to use AI to investigate a real infrastructure failure.
How this founder-led project changed my view of product work, assumptions, and the role of AI in execution.
The current public platform, learning surface, and public registration record can be viewed directly here.
Soulful Asia is current work, and the live official website shows the platform in its public operating state.