Case Study 01 Mobile / AI / EdTech

Mylo — an AI upgrade.

Redesigning a parenting app around a personalised AI assistant that pulls from real community wisdom, plus four supporting features that make day-to-day parenting feel less overwhelming.

Role
UX Research
UI Design
Team
Shrey Patel
Devanshu Gadani
Jesal Rathore
Timeline
14 Weeks
Year
2025
01 / Overview

A digital companion that understands, adapts, and evolves.

The Project

Mylo AI Assistant is a feature designed to enhance the Mylo parenting app by integrating an AI-powered companion. The feature, known as Mylo GPT, offers personalised support to young parents by drawing on data from Mylo's existing community forums and chats.

Alongside it, we introduced four supporting features: a sleep tracker, a baby profile page with growth stats and projections, a medical record repository, and a memorabilia section for capturing important moments.

The goal wasn't to bolt AI onto the app for novelty. It was to build a system that genuinely anticipates what a parent needs at 3am, when scrolling forums isn't an option.

02 / Problem

Information was abundant. Reassurance wasn't.

What we found

New parents face a peculiar kind of overwhelm. There is no shortage of information online — if anything, there is too much. The challenge isn't access, it's trust and time. Sifting through forums for advice that actually applies to your six-month-old at midnight is its own form of exhaustion.

We narrowed the problem to three pressure points:

01
Time-poor users, info-rich problem

Parents were using community forums but couldn't easily get to the answer they needed without reading through long threads.

02
Fragmented tracking

Sleep data, growth milestones, medical records, and memories all lived in separate apps, notebooks, or photo libraries.

03
Generic answers, specific lives

Generic parenting advice rarely matched the specific stage, temperament, or circumstance of the parent asking.

03 / Process

A five-step design process.

How we worked

We followed a user-centred design loop, beginning with deep research and ending with usability testing on high-fidelity prototypes. Each step fed into the next, with insights from interviews shaping every wireframe and screen we built.

01
Discover
User interviews, surveys, competitor analysis to surface real pain points.
02
Define
Personas, project goals, and design principles to guide every decision.
03
Ideate
Brainstorming features, sketching flows, and aligning on what to build first.
04
Prototype
Low-fidelity wireframes evolving into polished, interactive high-fidelity mockups.
05
Test
Usability testing with real users to validate, then iterate where needed.
04 / Research

We listened before we designed.

User Research

To understand the needs of young parents, we ran in-depth interviews with current Mylo users and analysed how they were using competing parenting apps and forums. We looked for patterns in what they reached for first, what they gave up on, and where the existing tools failed them.

Three insights came up consistently across every interview:

Insight 01
Real-time support matters most at 3am.
Parents wanted immediate, contextual answers when their baby was unwell or sleepless, not a forum search.
Insight 02
Community trust outweighs expert authority.
Parents trusted other parents who had been there over generic articles, even when the parent wasn't an expert.
Insight 03
Tracking is a chore without a payoff.
Logging milestones felt like work unless the data showed back something useful — patterns, predictions, or peace of mind.
Insight 04
Memories deserve a home, not a camera roll.
Parents wanted a deliberate space for milestones, separate from the chaos of their phone's photo library.
05 / Personas

Two parents, two realities.

Who we designed for

We built two primary personas to anchor design decisions. They captured the spectrum we kept seeing in interviews: from first-time parents in survival mode to seasoned ones who wanted control and continuity.

Persona 01 / Primary
Sarah
29 Marketing Manager New York
Stage
First-time mother of a six-month-old.
Needs
Quick, reliable advice on baby care, easy sleep tracking, and a way to document milestones without friction.
Pain points
Difficulty finding relevant advice quickly. Not enough time to read long forum threads.
Persona 02 / Secondary
Mark
33 Software Engineer San Francisco
Stage
Father of a one-year-old toddler.
Needs
Reliable info on growth and health, easy access to medical records, and a way to share important milestones with family.
Pain points
Frustration with fragmented tracking tools. Limited time to seek out advice when issues come up.
06 / Design

From wireframes to a cohesive system.

Design system

The visual identity carried Mylo's existing brand language — warm corals and sunlit yellows that already meant something to existing users — and tightened it into a system. We defined a typographic hierarchy using Montserrat and Gotham Rounded, an icon set, button states, and special interactive elements like the feedback slider, sleep tracker bars, and activity ring.

The result was a kit that any future feature could be built from without breaking the feel of the app.

Key Screens
01
Dashboard

The central hub showing upcoming milestones, events, and the baby's current age and weight. Designed to give parents a quick, calming overview.

02
Mylo GPT

An AI chat interface that pulls insights from real community posts to give personalised, grounded answers — not generic ones.

03
Sleep Tracker

A daily and nightly sleep log with visual patterns over time, helping parents spot trends without spreadsheet thinking.

04
Predictions & Events

A calendar of upcoming developmental milestones and important dates, so parents stay one step ahead.

05
Baby Profile

A space combining growth stats, saved documents, and a "Look Back" memorabilia section for milestone memories.

07 / Testing

Then we watched people use it.

What users said

We ran usability tests on the high-fidelity prototype with parents matching our personas. Most feedback was positive, but the most useful kind was the friction we hadn't expected.

Mylo GPT
Loved the personalisation, wanted follow-ups.
Users appreciated the real-time, community-grounded responses. Several asked for the ability to ask follow-up questions in-thread.
Sleep Tracker
Easy to log, easier to forget.
Logging was smooth, but parents asked for gentle reminders so they wouldn't break their tracking streak.
Baby Profile
Clear, but wanted more control.
Stats and projections were praised for clarity. Parents wanted more options to customise what was visible at a glance.
Memorabilia
Loved it. Wanted to share it.
Parents loved having a deliberate space for memories, but wanted to share milestones directly with family.
08 / Outcomes

A digital companion that actually companions.

What it became

Mylo GPT and the supporting features addressed the real needs surfaced in research. User satisfaction during testing rose noticeably, with positive responses to the personalised AI support, the ease of tracking sleep and milestones, and the sense of having a dedicated space for medical records.

The bigger outcome was strategic: Mylo positioned itself as a full-spectrum parenting tool rather than a single-purpose app, and we left the team with a foundation that future features could plug into without redesigning the system.

"

A great AI doesn't need a complicated interface. It needs to anticipate, learn, and stay out of the way.

Next Steps

From here, the work continues: refining Mylo GPT to handle multi-turn conversations, exploring personalised recommendations driven by sleep and growth data, and expanding the system as new feature areas open up.

Continuous user testing will keep the product honest. So will paying attention at 3am.

Next Project / 02

Roots research.

Read the case study