Iris
Overview
Iris is an AI-powered emotional check-in companion that helps users build self-awareness through daily mood tracking, pattern recognition, and conversational reflection. Designed in 24 hours at a designathon, Iris bridges the gap between journaling and therapy with a low-friction daily habit.
Problem Discovery
The Problem
Most people lack a consistent, low-effort practice for processing their emotions. Journaling feels like homework. Therapy is inaccessible. The gap between noticing something is wrong and getting support is too wide.
High friction check-ins
Existing mood apps require filling out lengthy forms that feel clinical rather than human.
No pattern visibility
Users can't connect today's mood to last week's — data is collected but never meaningfully surfaced.
Lack of agency
Apps tell users what to do. Iris reflects back what users already know — and helps them act on it.
Features
What Iris Does
Conversational Check-In
A 60-second AI-guided prompt flow that feels like texting a thoughtful friend, not filling out a form.
Mood Landscape
A visual map of emotional patterns over time, highlighting recurring triggers and positive anchors.
Reflective Nudges
Contextual reminders that reference past entries — "You felt this way last Tuesday too. Here's what helped."
Private by Default
All entries are stored locally. No data leaves the device without explicit user action.
Behind the Idea
Research
Given the 24-hour constraint, research was compressed into a rapid desk review of existing mental wellness apps paired with 3 guerrilla interviews conducted in the first two hours. Key insight: people want to feel heard, not tracked.
"I downloaded like six of these apps. I stopped using all of them after two weeks. They feel like chores."
— Guerrilla Interview, Participant 2Ideation
Exploring Solutions
With only 24 hours, ideation was time-boxed to 45 minutes. The team used a "Crazy 8s" sprint to generate interaction patterns, then voted on a conversational UI model that prioritised warmth and speed over completeness.
Low-Fidelity
Wireframes
Wireframes were sketched and digitised within 3 hours, focusing entirely on the core check-in loop. Secondary features were deferred to keep the prototype testable within the event's constraints.
Testing
Usability Testing
3 rapid prototype tests were run in the final 4 hours of the designathon. Participants interacted with a Figma prototype on a phone while thinking aloud.
Finding 01
The opening prompt was too abstract. Replaced "How are you?" with "What's on your mind right now?" — completion rate jumped.
Finding 02
Mood visualisation was confusing without a brief onboarding moment. Added a single tooltip on first use.
Finding 03
All 3 participants wanted to revisit previous entries mid-session — added a subtle "look back" entry point to the home screen.
High-Fidelity
Final Designs
The final UI uses a soft violet palette to evoke calm and introspection. Typography is oversized and generous — designed for late-night use when users are most emotionally available.