
UX/UI Design Student
January–June 2023
University of Washington P&CE
Figma, Adobe Photoshop,
Optimal Workshop
(Ideation phase) Nicole DesRoberts, Esther Cheng, Khadijah Al-Shami




"If I feel I'm struggling to connect, it feels like discordant and like trust is not present. And it feels disharmonious and like something I want to resolve."
"Listening before you speak... Listening is absolutely crucial."
From this, we created two user archetypes:


After selecting a direction, I mocked up some basic wireframes of screens displaying key features. This helped me get feedback from fellow students.

Feedback from concept testing was positive on the personalization and gamification elements, and some students saw potential in using AI to discover “unspoken rules,” but others were skeptical of AI chatbot technology’s capability to handle the task effectively.
"I feel like job interview politics that goes on that are just unspoken rules, right? And I feel like something like this really helps them to would it be so beneficial to like, learn these skills in a more relaxed environment?"
"I like it, but I feel as if it's putting a lot of trust on the Al in the development of the Al as opposed to you know, having a real person who I know has credentials."
"Most helpful would probably be the track performance over time. That's an interesting way of seeing feedback in order to improve Yeah, the feedback screen also does that but I like my graphs and charts."




.png)





Feature Creep
When I presented my final prototype, my course instructors said the app design was well-refined but too feature-heavy on the home screen.
They were right; I had envisioned the home screen as a catchall return point for every user flow that would redirect users into new flows, but I could have thought more about the experience of that page itself. Perhaps it could have been designed to resemble progress toward the user’s goals and immediate next steps.
Considerations for Human-Centered AI Implementation
This project did not include a specific exploration of prompt engineering for AI implementation, but that would be a necessary next step. AI technology, and my understanding of it, have advanced since this project was completed.
Some things are clear. A reliable prompt template would be needed to ensure chatbots adopt appropriate personas in conversation scenarios. Because users need to feel like they have permission to try new things and fail, AI should be prompted to be forgiving and encouraging in its feedback.