CHI 2025 Student Design Competition · AI · Maternal Health · User Research

MomSAFE

An entry in the CHI 2025 Student Design Competition: mixed-methods research with expectant mothers and OB/GYNs on where AI belongs in prenatal care, and where it doesn't. I translated the findings into a design direction for personalized, real-time pregnancy support.

My Role

UX Researcher

Team

3 members

Focus

AI-Driven Maternal Healthcare

Methods

Expert Interviews · User Research
Synthesis · Design Direction

Alignment

UN Sustainable
Development Goals

Project Resources

PDF

Research Paper

Full study & findings

→

PNG

Infographic

Visual research summary

→

At a glance

The challenge

Mothers couldn't parse medical language or tell a normal symptom from a serious one, and in one in three U.S. counties there is no obstetric clinician to ask.

What I did

As UX Researcher on a CHI 2025 Student Design Competition entry, I ran expert interviews with labor nurses and OB/GYNs alongside research with expectant mothers, then turned the findings into the design direction.

The outcome

A proof of concept for AI-assisted prenatal support in language mothers actually understand, with the trade-offs the research forced on us documented beside it.

Introduction

Empowering expectant mothers through AI-driven care.

An expectant mother told us: “I would love a chat feature with professionals at 2 AM when I'm struggling.” Pregnancy doesn't follow office hours, and for millions of women there is no one to ask when something feels wrong at midnight.

As the UX Researcher I ran expert interviews with labor nurses and OB/GYNs alongside research with expectant mothers. The gap wasn't only access to doctors, it was access to understanding. Mothers couldn't parse medical language or tell when a symptom was normal.

MomSAFE uses LLMs and machine learning to bridge that gap: synthesizing user inputs and clinical data to deliver personalized, real-time pregnancy insights in language mothers actually understand.

My role: UX researcher on a 3-person team. I conducted interviews with expectant mothers and healthcare providers, synthesized findings into design requirements, and shaped the research direction and recommendations.

Global Alignment

Designing for the UN's Sustainable Development Goals.

Goal 3

Good Health & Well-being

Ensuring access to essential maternal healthcare for all women regardless of location or socioeconomic status.

Goal 5

Gender Equality

Empowering women with personalized, evidence-based care and the information they need to make informed decisions.

Goal 10

Reduced Inequalities

Closing the healthcare gap in underserved communities where access to OB/GYNs and prenatal care is limited.

Problem Statement

The maternal healthcare gap is a crisis hiding in plain sight.

One in three U.S. counties is a maternity care desert, with no birthing facility and no obstetric clinician (March of Dimes, 2026), and more than 700 women die every day worldwide from preventable pregnancy and childbirth causes (WHO, 2023). Yet most mothers we asked called their care “fine.” They had no benchmark for what adequate looks like.

1 in 3

of U.S. counties are maternity care
deserts: no birthing facility, no
obstetric clinician (March of Dimes, 2026)

700+

women die every day worldwide
from preventable pregnancy and
childbirth causes (WHO, 2023)

1 in 4

of U.S. pregnancies do not get
early and adequate prenatal
care (Healthy People 2030, 2024)

User Research & Insights

Doctors and mothers told us
completely different stories.

Healthcare professionals saw a literacy problem: mothers panic over normal symptoms or ignore serious ones because the terminology is opaque. Mothers saw an access problem: they knew something felt off but couldn't reach anyone when it mattered. Both were right.

Expert Consultations

Healthcare Professionals

Consulted with labor nurses and OB/GYNs who advocated for pain-rating scales and simplified medical terminology to make clinical information accessible to all mothers.

User Interviews

Expectant Mothers

Interviewed pregnant women who expressed a strong need for real-time support, especially during off-hours when their providers aren't available.

It's not expected that an average mother understands complex medical terms.

Labor Nurse

Expert Consultation

I would love a chat feature with professionals at 2 AM when I'm struggling.

Expectant Mother

User Interview

Key Features

Every feature traces back to
something a mother or nurse told us.

The chatbot exists because mothers need answers at 2 AM. Simplified language exists because a labor nurse told us “it's not expected that an average mother understands complex medical terms.” Check-ins exist because OB/GYNs called longitudinal data the single most valuable input for early risk detection.

AI Chatbot

24/7 multilingual support for maternal health inquiries: answering questions, providing guidance, and connecting mothers to resources any time of day.

Survey Check-ins

Regular symptom tracking for mood, physical health, and pregnancy milestones to build a longitudinal view of each mother's journey.

Appointment Booking

Direct access to healthcare providers with streamlined scheduling to reduce the friction between recognizing a concern and getting professional care.

Data-Driven Insights

AI-powered trend analysis that surfaces patterns across check-ins; proactively flags potential risks before they become emergencies.

Privacy & Security

User-controlled data with transparent policies; trust is non-negotiable when it comes to sensitive health information. Every mother decides what to share and with whom.

Design Direction

Research-informed principles for accessibility, inclusivity, and trust.

A striking finding: health apps often made mothers more anxious. Overloaded dashboards, alarming colors and jargon turned tracking into stress. That shaped the design: surface only what's relevant, frame insights as guidance rather than warnings, explain in plain language.

Accessibility

Simplified language, multilingual support, and intuitive navigation for mothers of all backgrounds and tech comfort levels.

Inclusivity

Designed to serve all mothers, regardless of location, language, or access to traditional healthcare resources.

Trust

Evidence-based recommendations, transparent AI, and user-controlled data create a safe space for health decisions.

Design Decisions

Two trade-offs the research forced
us to make.

Mothers and providers gave us conflicting signals about what MomSAFE should be. Each conflict surfaced a real design choice, and the wrong call in either direction would have undone the research.

Reactive logging, or longitudinal check-ins?

Mothers wanted real-time help at 2 AM. The obvious response was a reactive symptom logger. OB/GYNs pushed back: reactive data is the least useful kind. They needed longitudinal patterns to catch the early signals that distinguish normal pregnancy variation from a developing complication. I carried both into the concept, with longitudinal check-ins as the spine and the AI chatbot pulling from that same context.

More upfront commitment from mothers (daily input vs. only-when-needed). Mitigated by keeping each check-in under 30 seconds.

Clinical accuracy, or lay language?

A labor nurse summarized the problem: "It's not expected that an average mother understands complex medical terms." But pregnancy is high-stakes, and misstating a clinical concept could cost real safety. The reframe: language isn't accuracy vs. accessibility; it's about who it's for. Plain language defaults for mothers, full clinical language for provider-facing surfaces, tappable definitions in between.

Two parallel content systems to maintain. Accepted because forcing mothers to learn medical terminology to use a maternal health app was exactly the gap MomSAFE existed to close.

Impact & Future Vision

Where MomSAFE goes from here.

A proof of concept with a solid research foundation, and a clear list of what it would still need.

MomSAFE is a proof of concept. What we showed is that AI, plain medical language and 24/7 access address pain points mothers and providers independently confirmed. Going further would need clinical validation, regulatory review and healthcare partnerships.

Telehealth Integration

Direct doctor consultations built into the platform to bridge the gap between expectant mothers and healthcare professionals.

Predictive AI Diagnostics

Real-time AI-powered diagnostics to detect high-risk pregnancies earlier, when intervention matters most.

Global Language Support

Expanded multilingual capabilities for a truly inclusive experience: ensuring no mother is left behind because of a language barrier.

Thanks for reading

Let’s talk.

My one-page resume, or reach me directly.

Resume (PDF) LinkedIn Email