Case Study AI Assistant
Project Summary
I served as lead designer and Agile coach for the development team. We built a Vue-based AI assistant that provided real-time suggestions and resources to contact center agents during live chats. Agents were able to respond more quickly and accurately, while reducing customer chat wait times.
Key Highlights
- Led a cross-functional team of 6 to deliver a fully functional AI assistant in 3 months
- Led contextual interviews and usability testing with contact center agents to uncover pain points and validate early prototypes
- Designed an intuitive interface that integrated seamlessly with existing chat workflows
- Implemented Agile methodologies to ensure rapid iteration and feedback loops
- Guided discussions on AI limitations and ethical prompts to ensure the assistant supported (rather than replaced) human decision-making
Process
We began with user research; interviewing and observing contact center agents to understand their challenges and workflows. I facilitated collaborative workshops to define personas and usage scenarios, which shaped our design priorities. We developed low-fidelity wireframes and interactive prototypes, refining them through multiple rounds of usability testing.
Alongside design work, I coached the team on Agile practices, helping them plan in sprints, prioritize based on user impact, and reflect regularly in retrospectives. We adjusted our approach in real time, staying flexible as new needs emerged.
Results
The assistant led to measurable gains in both agent speed and confidence. Average chat wait times dropped, and agents reported greater satisfaction with their tools. Stakeholders recognized the product for its practical value and its grounding in thoughtful, user-centered design. The project also served as a template for future AI-supported workflows within the company.