I designed, built, and deployed a fully custom research platform — web application, data transformation model, and analytics dashboard — for a US research organisation in under a week, managing ~9,000 US towns, producing 170,000+ clean records, at 3x less than traditional development.
Brief to deployed
< 1 wk
From brief to fully deployed solution — days, not months.
Cost vs. traditional build
3x
Cheaper than traditional custom development.
Clean records produced
170,000+
Analysis-ready records across ~9,000 US towns managed through the platform.
A US research organisation in the non-profit sector ran a large-scale, recurring survey programme that had outgrown its tools. The programme covered nearly 9,000 US towns and hundreds of charities — far more than its existing platform, Qualtrics, was designed to handle. The team had the data; they just could not get to it fast enough, and scaling the programme without everything breaking was a constant risk.
Every survey wave, the same problems repeated:
The solution started with deep knowledge of Qualtrics and a clear understanding of exactly what the research team needed. I then used AI to design, build, and deploy a fully custom solution in a fraction of the time and cost of traditional development. The build followed three steps:
That produced three custom deliverables, all fully owned by the client:
The platform went from brief to fully deployed solution in under a week. The AI-assisted build was 3x cheaper than traditional custom development. It now manages ~9,000 US towns and has produced 170,000+ clean, analysis-ready records.
Most importantly, the research team can focus entirely on the findings. They have ready-to-export customised reports at town, charity, or state level in minutes, with no data preparation required. They manage the full survey programme through a simple browser interface built for their exact workflow. And they own a flexible, adaptable system they can extend as the programme grows — without needing a developer on call.
The speed came from combining three things: deep knowledge of Qualtrics, a clear understanding of exactly what the research team needed, and AI-assisted development. I first mapped the full survey workflow — how towns were assigned, how charities were randomised, and how data flowed from respondent to analyst — with no assumptions, then used AI to design, build, and deploy the solution in days rather than months.
The client fully owns the code and can adapt it as the programme evolves. There is no vendor lock-in, no licence fees, and no waiting on a feature request — the organisation can extend the system as the programme grows without needing a developer on call.
The AI-assisted approach made the build 3x cheaper than traditional custom development, and it was delivered in a fraction of the time — under a week from brief to fully deployed solution instead of a months-long development cycle.
Three custom deliverables: a purpose-built web application for managing the survey programme, designed around how the team actually works; a repeatable data transformation model that automatically cleans and structures raw survey exports every wave; and a fully customised qualitative analytics dashboard the client can filter at state, town, or charity level to answer any research question instantly.