Designing a Smart Mirror for Postpartum Body Image

Reflect was my HCI iSchool capstone, a project my team ran end to end, from problem to working prototype, inside an agile process where we each owned a role. I was one of two product designers, and together we researched, designed, and built the product.

What I owned: the vision and the human-mirror interaction, meaning how someone actually uses a mirror, from how they control it to how a session unfolds. My co-designer led more of the visual UI.

The premise: dissatisfaction with one's body is common, and it intensifies during pregnancy and the postpartum period. We built Reflect, a smart mirror that turns a daily, ordinary moment, looking in the mirror, into a small act of wellbeing. It blends daily check-ins, intention setting, breathing exercises, reflection prompts, and clinically grounded mirror exposure therapy to help users slow down and develop a kinder relationship with themselves.

The work won the Design Award at our 2023 Capstone Night.

We started capstone with no project and one shared goal: make something that matters. As we traded the issues we cared about, body image kept surfacing, and the numbers made it impossible to ignore.

A literature review led us to a clinical practice most people have never heard of: mirror exposure therapy.

It's a cognitive-behavioral technique for body dysmorphia. Users gradually increase time spent observing their reflection to confront negative body-image perceptions and reduce avoidance, while a guided process helps them challenge distorted thoughts and build more realistic, positive self-perceptions. Coping techniques like deep breathing and mindfulness manage anxiety during exposure.

It works, but it lives almost entirely inside therapy rooms, and the normalization of body dysmorphia keeps people from seeking it out.

The same review told us our audience was a good fit for this approach. Pregnant women strongly prefer talk-based support over medication for depression², and mindfulness-integrated CBT has been shown to reduce depression and anxiety during pregnancy³.

Then we scanned what already existed: self-help books, apps, and fitness mirrors. Nothing combined mindfulness and mirror exposure therapy in someone's own home. That gap became our opportunity: bring a clinically grounded practice into the home, in a form people would actually use every day.

We focused on the people most disproportionately affected by negative body image during a specific life stage.

Body image during this period is tied to emotional well-being, mental health, maternal health, and bonding with the baby.

To represent our audience we built two flexible user types, pregnant women and postpartum women, that could stretch to cover a broad population. They were grounded in real conversations: we interviewed our own mothers and a friend who was pregnant at the time, and the types kept evolving as our understanding deepened. Shared wants centered on self-care, support, alone time, and organization; shared pain points included limited mobility, body change, morning sickness, anxiety, and a desire for community and better body perception.

To align our professor, peers, and team on a shared vision, we storyboarded the experience step by step, mapping exactly how someone would move through a session.

This was the part I owned. The hardest design question was: how should a person interact with a mirror? We weighed three models:

To understand smart-mirror behavior firsthand, we went to Lululemon to try theirs in person, testing colors, interactions, and the overall design language. It directly shaped our UI decisions:

• White text is easiest to read over a mirror reflection.
• Bright colors pop most against the reflection.
• People read a mirror top to bottom, so hierarchy should be vertical.
• A word-forward design language reads better than icons.

We built our first high-fidelity prototype, then put it in front of users. Four findings reshaped it:

We codified everything into a design style guide and system, then built Reflect around five core features:

View the Figma prototype →
Watch the video demo →

Every instinct I had as a designer said to reduce friction. Reflect is the one product I've worked on where the friction was the point. When testing showed people racing through the prompts, the fix was a timer that refused to let them move on. Mirror exposure therapy only works if you stay with the discomfort, so the interface's job was to slow someone down.

That reframed how I thought about the whole product. We were taking a clinical practice out of a therapist's office and putting it in someone's bedroom, during a vulnerable stretch of their life. It's also why choosing the less elegant input, touch over voice, was the right call.

What I carry from Reflect is the habit of asking how a design makes someone feel while they use it.

If we took Reflect further, we'd:

• Add progress analysis so users can see their journey over time
• Expand the product scope
• Run more usability testing
• Improve the efficiency of the program
• Add more features