MamaFlow

The idea behind MamaFlow started from a simple insight: new mothers need support, but rarely have the time or hands free to seek it. The early months of motherhood are filled with constant change — emotional highs and lows, disrupted sleep, and new responsibilities — leaving little room for self-care or meaningful connection.

From this reality came the key design principles that shaped every decision in the project:

• One-hand usability — the interface had to be navigable while holding a baby or stroller, with large touch zones and simple gestures.
• On-the-go accessibility — features needed to deliver value in short, interrupted moments rather than requiring long attention spans.
• Voice and audio-first options — because many mothers can listen when they can’t read or type, audio content and voice input became essential interaction modes.
• Emotional clarity — visuals had to feel calm and reassuring, avoiding overstimulation through minimal palettes, soft gradients, and gentle typography.

These principles guided my work with UXPilot, where each screen was created and refined through text-based prompts. Designing this way — my first-ever high-fidelity UI experience — required precision in language and intention. Every prompt became a design decision, defining how warmth, calmness, and usability should manifest visually.

My approach

Because the UXPilot competition offered only a limited number of credits, I decided to take on an extra challenge — to complete the entire project using only the basic credits provided, without purchasing additional ones. I wanted to see how far I could go with clear thinking, structured prompting, and no prior experience in high-fidelity UI design.

This constraint pushed me to make every prompt count. It meant crafting precise, strategic instructions — and iterating thoughtfully rather than endlessly. To achieve that, I collaborated closely with ChatGPT, refining each prompt to capture not only the functional requirements but also the emotional tone of the design.

Challenges and key learnings

Working with UXPilot was both exciting and informative. As someone creating high-fidelity UI for the first time, I quickly learned that designing with AI isn’t just about what you want — it’s about how precisely you describe it. To start, I used the example prompt provided by UXPilot as a reference, and it worked well as a foundation for everything I needed to include in my own prompts. Of course I then needed to specify a lot of details which did not come to my mind at first.

One of my biggest challenges was color definition. At first, I assumed I could simply write “soft cream” or “warm lavender,” expecting the AI to interpret the tone visually and use it cohesively. It didn’t. The system rendered inconsistent gradients or unexpected color temperatures. Eventually, I discovered the most reliable approach was to specify both a reference to an existing screen and the exact hex code — for example,

“Use the same background as on the homepage (solid warm cream #FFF8F3).”

That small line changed everything — ensuring visual continuity across all ten screens and teaching me the importance of linguistic precision in AI-driven design.

Another early mistake was assuming the layout structure could be adjusted later. In UXPilot, the layout must be defined within the prompt itself. My first two attempts failed completely — I lost 36 credits just trying to “fix” something that could have been avoided with clearer upfront instructions. It was a frustrating but eye-opening realization: the prompt is the blueprint. Once I started thinking like that, the process became far more efficient and intentional.

And eventually there was another challenge for me with the bottom navigation bar. My first attempt to define it in the prompt looked like this:

“Bottom Navigation Bar: Fixed global navigation with 4 icons: Home, Community, Care, Food. Rounded active indicator, easy reach for thumb.”

But what I got was basically a navigation bar from last century. So I asked ChatGPT what I could do to make it look more modern. I got the following answer. And you know what? It worked just right!

“Bottom Navigation Bar:

Fixed global navigation with 4 rounded icons and labels.

Items:

1. Home 🏠 – Returns to main dashboard.

2. Community 🤱 – Opens "In this together" for sharing and listening.

3. Care 🌿 – Opens "Breathe" section with quick self-care audio.

4. Food 🍽️ – Opens "What we eat" for recipes and meal ideas.

Icons should be rounded, filled vector icons with muted colors and short text labels below.

Active icon highlighted by soft background circle and gentle shadow for easy thumb recognition.”

These moments reinforced one of my biggest takeaways from this project: AI is not a shortcut — it’s a dialogue. The quality of the design directly depends on the clarity, structure, and empathy of the language you use to describe it.

The 10 screens

1. Homepage – Central dashboard introducing the app’s key sections: Breathe, Talk to Someone, Community, and What We Eat.
2. Community (In This Together) – Discussion area for connecting with other moms, sharing advice, and supporting each other.
3. Breathe (Care) – Quick self-care screen with short breathing or relaxation exercises, designed for one-hand use and audio guidance.
4. Food: What We Eat – Overview of family meal ideas and recipes suitable for different needs (baby-friendly, dairy-free, etc.).
5. Talk to Someone – A space for emotional support with mood check-ins and access to professional help or peer connections.
6. Find Support – Directory of experts, hotlines, and nearby moms open to sharing their experiences.
7. Daily Reflection – Mood tracker with weekly and monthly visualization to support mindfulness and emotional awareness.
8. Recipe Detail – Detailed recipe screen with ingredients, preparation steps, and quick actions for saving or sharing.
9. My Space – Personal dashboard with user’s saved reflections, favorite recipes and progress overview.
10. Settings – Customization hub where users can manage notifications, language, theme mode, and personal preferences

Conclusion

In the end, the combination of UXPilot and ChatGPT allowed me to create something I genuinely consider effective — all within just two naps of my baby. It made me realize that with the right mix of smart tools and clear intent, a little bit of magic can happen even without prior experience.