HelpMum develops AI-powered maternal and child healthcare tools to improve maternal and infant healthcare in underserved communities. HelpMum MamaBot AI and HelpMum VaxAI provide pregnancy and vaccination support through WhatsApp, while HelpMum StratifyAI helps health workers detect high-risk pregnancies early. Together, these innovations create a continuous care pathway from pregnancy to childhood immunization using accessible, low-cost digital technology.  

This project reflects values of ethical AI, digital innovation, and equitable healthcare access. It combines AI-driven pregnancy risk assessment with evidence-based maternal health support to help healthcare workers identify complications early and improve referral systems. Through accessible WhatsApp-based chatbots, the project expands low-cost healthcare access for underserved communities, aligning with the integration of technology, research, and public health to improve outcomes. 

Use of AI  
HelpMum integrates AI across its maternal and child healthcare innovations. MamaBot and VaxAI use Large Language Model–based conversational AI to provide caregivers with trusted pregnancy and vaccination information through WhatsApp. StratifyAI uses predictive machine learning and AI-powered clinical decision support to help health workers identify pregnancy risks early and provide tailored care and referral recommendations.  

Software 
StratifyAI is powered by a Python FastAPI backend hosting a proprietary pregnancy risk-stratification model that helps healthcare workers assess pregnant women and generate risk-informed recommendations. The frontend is built with React.js, providing dashboards for real-time risk visualization and clinical decision support. HelpMum’s WhatsApp chatbots VaxBot, VaxAI, MamaBot, and MamaBotAI are built on Turn.io to support large-scale maternal health and immunization engagement. The system uses secure API integrations, PostgreSQL for clinical data storage, and AI-powered conversational support to deliver scalable, accessible, and evidence-based maternal and child healthcare information. 

Hardware 
Our innovation uses cloud infrastructure and dedicated development hardware to support deployment, scalability, and testing. AWS CloudFront enables low-latency access to the platform and AI APIs across regions, while AWS GPU-powered G-Series instances support real-time AI processing for the StratifyAI model and conversational systems. AWS S3 is used for secure storage of clinical datasets and encrypted user data. Development and testing are carried out using Apple MacBook Pro (M3) workstations, Docker environments, and Figma for UI/UX design. The platform is also tested across low-end and mid-range Android and iOS devices to ensure accessibility and usability in resource-constrained communities. 

HelpMum (NG) was founded by Dr. Abiodun Adereni to improve maternal health through low-cost digital innovations. His work includes collaboration with Google Research and Vanderbilt University on the award-winning AI tool, ADVISER, focused on improving vaccination uptake. In 2023, he became an Ashoka Fellow. HelpMum has received support from Google, Meta, Merck, Gavi, and the Patrick J. McGovern Foundation. Dr. Adereni is also a Youth Advisor for the World Bank and holds a Doctor of Veterinary Medicine degree from the University of Ibadan.  

Maternal mortality remains one of the most persistent and preventable crises in global health. In many of the contexts where this project operates, a pregnant woman’s risk of dying in childbirth is dramatically higher than in high-income countries, not because medical knowledge does not exist, but because it does not reach her in time, or at all. 
 
This project receives the Grand Prize for the clarity and urgency of its mission, and for the seriousness with which that mission has been pursued. The use of AI for early pregnancy risk stratification, identifying women at heightened risk of life-threatening complications, is not new in principle. What distinguishes this work is its grounding in the operational realities of low-resource health systems: limited connectivity, understaffed clinics, and communities in which institutional trust must be earned rather than assumed. 
 
The jury was particularly impressed by the project’s approach to AI ethics. The team provided substantial documentation of its data governance framework, including how patient data is collected, anonymized, protected, and potentially shared with third parties. While we noted some gaps, particularly regarding escalation protocols in the event of serious system failure, the overall ethical architecture was judged to be robust and thoughtful. In the realm of health data, among the most sensitive categories of data that exist, this level of rigor sets an important benchmark for future applicants. 
 
This is a project that addresses a fundamental human need with humility, precision, and care. It merits the jury’s highest recognition.