Context & My Role
KIMI was a human-centered AI course project developed by six design and computer science students. Starting from a fictional brief by Ubiquitos AI AB, we explored how a wearable organiser could help first-time parents share everyday information and reduce mental load.
As project leader, I coordinated the group, organised meetings and helped keep our work aligned with the timeline and goals. I also created the concept map and future storyboard shown on this page, translating our ideas into visual scenarios.
The open brief gave us responsibility for defining both the users and the purpose. Our question became: how could AI support coordination between parents without adding more screen time or taking over decisions that belong to them?
Understanding Users
Our team gathered 59 questionnaire responses and conducted eight semi-structured interviews. We used personas, storyboards and an emotional journey to translate the findings into design goals rather than starting with the technology's possibilities.
The research pointed to an imbalance in everyday planning: the parent spending more time at home often held most of the information about routines, appointments and tasks. Short updates between parents could leave gaps, while recording activities added another responsibility.
These insights guided us towards shared awareness and hands-free interaction. We focused on making it easier to keep each other informed, rather than using AI to give advice about a child's needs.
Turning Insights into a Concept
We developed KIMI as a voice-driven organiser worn on clothing. Parents could log an activity, add a note or record a task through speech, with information available through a companion calendar interface.
My concept map made the relationship between the badge and both parents' phones visible. In my future storyboard, I explored how these interactions could fit into everyday family life, helping communicate the intended experience beyond individual screens.
The team developed wireframes, a Figma prototype and a physical badge prototype. The badge concept combined a microphone, speaker, power button, magnetic attachment and an illuminated edge, using light to communicate states such as listening or confirming a logged activity.
The badge concept
Light as feedback
Animation of the badge's edge light changing colour, exploring how light could signal the badge's state.
Testing & Iteration
In an early Wizard-of-Oz test, a participant tried logging a nap and adding a note while our team simulated the system's responses. This let us explore the interaction before every technical part was in place.
The participant preferred speaking without pressing a button, which informed the wake-word concept. Later tests raised questions about the light's placement and the badge's size. The feedback showed us that clear interaction signals needed to be considered alongside wearability and baby safety.
The team also developed a working demonstration using n8n, Ollama and Google Calendar, converting speech into a command that triggered a calendar action. This was narrower than the visual prototype: shared wearable pairing, personalised predictions and further integrations remained future ideas, not completed features.
Physical prototype
Future storyboard
Leading & Collaborating
Leading an interdisciplinary group meant coordinating different programmes and parallel course schedules. I organised meetings and supported planning through online communication and a clear division of tasks, helping the group stay aligned as the work progressed.
Working alongside computer science students strengthened my patience and understanding of technical development. I gained a clearer sense of the time and structure needed when code, data and troubleshooting come together.
A key lesson for me was the importance of a shared language between designers and developers. Visual scenarios helped make the intended experience concrete, while collaboration made the difference between a proposed interaction and a demonstrated capability clearer.
Key Takeaways & AI Ethics
This project reinforced the value of starting with users' needs and questioning where AI genuinely helps. Choosing organisation over advice about a child's needs gave the concept a clearer purpose and kept decisions with the parents.
Voice privacy, sensitive family information and misinterpreted commands also shaped our discussions. We explored local AI processing to give users more control over voice data, but external services and information sharing would still need safeguards. Our prototype was not a production-ready service with a complete data-governance framework.
I take away a stronger understanding of interdisciplinary teamwork, the value of testing interactions early and the responsibility to distinguish a future vision from what has actually been built. My role combined coordination with visual storytelling, helping connect our research, concept and intended experience.
Project team
Maria Fallström · Emma Schatz · Maddi Xu Persona Huici · Habib Allah Al Berotiy · Daniel Hopsch · John Widén Nordlund