Gabriella Ugwonali

KYRO

Role: Design Researcher & Engineer

02 / The problem

Efficiency Requires Emotional

Most scheduling tools track tasks. None of them track how those tasks make you feel. Our research found that people rarely notice the emotional cost of their day building up — they only recognize burnout in hindsight, once it’s already arrived.

Human proof point

One interview that shaped this most was with a mother of three watching her children struggle to stay on top of responsibilities while quietly carrying the same emotional weight herself. Another was with a Venezuelan Doordash driver describing the pressure of keeping up just to get by. Across nine interviews, the same pattern surfaced: getting things done was more than just logistics.

Space for visual: emotional cost map / journey diagram

Add diagrams, interview artifacts, or a burnout-before-after model here.

03 / Research & discovery

Nine conversations, one shared thread.

We conducted nine in-depth interviews across a wide range of ages and backgrounds. I personally led six of them, and worked closely with our other designer to synthesize the findings into empathy maps.

Productivity is emotional, not just logistical.

Structure and ritual often double as mood regulation.

“Wasted time” reads as a threat to identity, not just inefficiency.

Space for empathy maps and affinity synthesis

Honest limitation

Our sample skewed toward participants from more affluent backgrounds, given our recruiting location and timing — a limitation we’d want to address in a broader research pass.

04 / Points of view & ideation

From empathy maps to one clear direction.

We distilled our research into three core points of view, then generated 99 How Might We statements, narrowing 97 possible solutions down to three experience prototypes. Being closest to the empathy maps, I was one of the strongest voices pushing for the direction we ultimately pursued, an emotional check-in that helps people notice their patterns after each event in their day, rather than a passive tracker.

Representative POV

Lily was a 20-year-old highly organized Stanford student balancing an intense workload while protecting time for rest and connection. She needs planning support that respects both ambition and recovery, because staying productive should not mean ignoring the emotional cost of her week.

How might we help people notice emotional patterns without making reflection feel like homework?

How might we make planning feel adaptive instead of punitive?

How might we surface emotional insight after each event without interrupting the day?

99 HMWs → 97 solutions → 3 prototypes

05 / Design evolution

Designing something that feels like it’s listening.

I owned Kyro’s visual design language from the start. Our interviews kept surfacing a quiet desire to return to something more analog, so I began experimenting with physical paper textures to make the product feel more tangible inside a digital interface. I layered in soft gradients and watercolor-like transitions so the system could communicate state without feeling intrusive or alarming.

I owned Kyro’s visual design language from the start. Our interviews kept surfacing a desire to return to something tangible, so I began experimenting with physical paper textures to make the product feel more analog inside a digital interface. I layered in soft gradients and watercolor-like transitions so the system could communicate state without feeling intrusive or alarming. Through Iterative development,

Initial low-fi

Wireframes / first interaction model

Medium-fidelity Figma

Paper textures / early gradient language

Final high-fidelity

Final UI screens / motion states

Heuristic evaluation

34 total usability issues surfaced, 12 of them major. Each one fed directly into the next design iteration — from adding loading feedback during AI processing to preserving user input when switching between voice and text.

Mindfulness

Intuitive Design

Emotional Wellbeing

Responsible AI Use

Emotional, but intentionally non-sentient.

06 / Final prototype & technical implementation

Bringing the design system to life.

Alongside design, I worked on the technical build. I engineered the prompts for Google Gemini so the AI would generate warm, specific insights from spoken or written reflections without giving advice, making assumptions, or overstepping emotionally. I also built the Supabase database and the full onboarding flow, including OAuth login and a personalized welcome experience, while also working to keep the front-end code consistent with the design system throughout.

Space for onboarding flow, Supabase schema, or prototype architecture

Gemini prompt rules

Return only a JSON array. Generate exactly 3 insights. Each insight must begin with “You…”. Reference one specific detail from the user’s reflection. Do not ask questions. Do not give advice. Do not make assumptions. Keep the tone warm, specific, and emotionally careful.

Transparency note

Most of the product’s dialogue, weekly insights, and Kyro Scores were intentionally hard-coded for testing stability. The live chat insights were the one feature genuinely powered by the Gemini API in real time.

07 / Reflection & what’s next

What this project taught me.

Combining lanes

Most of my past experience kept computer science, visual design, and user research in separate lanes. This project let me combine all three and gave me a real sense of ownership, because I could trace the reasoning behind a decision from research all the way through to shipped code.

Real ownership

It was more demanding than I expected for a class project. I spent Halloween weekend building out user flows instead of going to parties, and Thanksgiving break mapping out system logic. The project required many late night phone calls with team members , butt the excitement never faded, especially watching real users respond to what we were building.

Continuing beyond class

I have been continuing working on the backend bringing it closer to production, with next steps including deeper pattern analysis, adaptive planning intelligence, and stronger accessibility support. Although there is no plan to go to market, it has been an incredible learning experience in design engineering.