The Collaboratory serves as the vital record for the higher education community engagement. However, the platform’s success relies entirely on the quality and quantity of data inputted by student workers, a process known internally as "Proxying." I did it myself and realized we weren't asking students to be researchers; we were asking them to be human bridge-builders for data that was already digital. I took the initiative to pivot my focus. Instead of just tweaking the existing forms, I decided to bring the tool to the data. I proposed a shift in our technical strategy: The AI Proxying Plugin
Impact
,
+
ASU collaboration with community database
Saving
sec
Average processing time
Student Workers
+
Students are first point of contact
Design Approach
I took ownership of the entire lifecycle of this feature. I advocated for a "Review-First" philosophy. I moved the user from a state of searching to a state of validating.
The Audit
I shadowed student workers to map every "click" and "tab-switch" in their 20-minute cycle.
The Prototype
I built a functional browser extension that could "scrape" the active tab and communicate with an LLM.
The Prompt Engineering
I spent weeks refining how the AI interpreted "regulatory jargon" to ensure it wasn't just grabbing text, but actually answering the specific questions required by database.
Key features
We built a system that doesn't just "talk" to you, but actually builds with you. We turned raw LLM data into structured, interactive components that feel like a conversation with a local friend.
Each AI-generated response references supporting website content to ensure transparency and verification confidence, while analysis occurs independently of user activity so workflows continue uninterrupted.
Tradeoff: Adds interface and technical complexity but improves trust, verification, and workflow efficiency.
