Tools: AI-assisted web design, rapid prototyping, adaptive content design
Audience: Online university call center agents
Challenge: Call center agents were misstating a new exam-break policy, leading to a 10% error rate and $60 fee waivers per mistake. Existing self-paced e-learning was failing to train agents sufficiently because it didn't account for the stress of a live call.
Solution: A learning ecosystem that renders the same instructional content into three distinct interfaces for different "moments of need".
Key Features:
A color-coded printed at-desk job aid that can be displayed at the agent's workspace for just-in-time performance support in the flow of work and during live-calls (5-second visual scan time).
A colorful, interactive, click-through decision tree that guides the agent through complex ADA and policy criteria.
Safe-to-Fail Sandbox (below): An interactive simulation for agents to practice scripts and fail safely before talking to students by experimenting with a variety of possible scenarios that might arise during a call.
Why Code-Based? Vibe coding allowed for a lightweight, responsive ecosystem that supports the agent from training to their thousandth call, prioritizing muscle memory over static knowledge.