Team: Women Genital and Fertility Health foundation (WOGAFH) | Fatimoh Adeoye-Adeniyi, Victor Adebayo, Comfort Bamikefa, Kawthar Hamod, Dr Yekini Ismail Damilare
Every consultation begins with a story. A woman sits on a wooden bench in a community health centre in southwestern Nigeria, waiting to explain something her body has been telling her for weeks: a cycle that has changed, a pain that returns, a pregnancy that has not come. When her turn arrives, she speaks in Hausa, the language in which she first learned to name her body and its rhythms. The nurse across from her listens closely and understands little, since her own languages are Yoruba and English. A colleague is sent to find someone who can translate. In the minutes that follow, the most private account a woman can give passes through a third voice, and the nurse is left to trust that its meaning has travelled intact.
Medical anthropologists have long observed that care depends on being understood, and that illness is lived and told in the language of everyday life before it is ever written in the language of a chart. In the local government area where the Women Genital and Fertility Health Foundation (WOGAFH) works, around 40% of residents speak Hausa, and nurses report that Hausa-speaking women make up a large share of those who come to them for reproductive health care. For these women, the distance between what they feel and what is heard can shape the care they receive.
WOGAFH, a foundation grounded in years of work on reproductive and fertility health, set out to close that distance with a tool that allows women and their health workers to understand one another directly. The team is building it as part of the Gender & AI Innovation Collective (GAIC) prototyping cohort convened by the A+ Alliance, drawing on the Feminist AI Sandbox developed by Women at the Table and Code for Africa.
Listening to the people at the heart of care
WOGAFH began by asking health workers what they needed. The team visited three community health centres and spoke with community health extension workers and nurses, who told them that language, more than any other barrier, stood between Hausa-speaking women and the care they came for. The team had first considered building in Pidgin, which reaches many people across Nigeria, and changed course once health workers made clear where the real gap lay. Odeda Local government of Ogun State population data confirmed what nurses described, giving the team a clear reason to begin with Hausa.
To capture how women actually describe their symptoms, the team organised outreach days at two health centres, where Hausa-speaking patients, with their consent and with a translator present, recorded the ways they explained their concerns. Some women were wary at first, and it was a trusted health worker who explained the purpose of the project and helped them feel at ease. The team had initially planned to visit women at home, and chose instead to meet them at the clinics they already know, which proved safer for everyone and kept the work moving.
Many families in the area turn to traditional medicine for reproductive health concerns, often before or alongside a visit to the health centre. WOGAFH recognised that a tool built only on clinical guidelines would miss a large part of how women seek care. The team also visited traditional treatment practitioners and, with their permission, recorded how they treat several common female reproductive health conditions.
This raised a question the team continues to think through carefully. Much published research on medicinal plants, including studies on conditions such as PCOS, anovulation and infertility, remains at the laboratory stage and has rarely reached clinical trials. WOGAFH is drawing on peer-reviewed research where results are significant and relevant, and is creating its own documented record of healers’ knowledge, a body of understanding that has rarely been captured in a form a machine can learn from. The team’s knowledge base therefore brings together Nigeria’s national Standard Treatment Guidelines, the standards that guide community health extension workers and nurses, World Health Organization guidance on the ethical use of AI in health, standard patient intake questions, and carefully selected research and testimony on traditional practice.
SmartHealth Community: How the tool works
SmartHealth Community has two faces. The first is designed for health workers. When a Hausa-speaking woman arrives, the health worker can hand her the phone so she can speak in her own language. The system transcribes and interprets what she says, and the health worker receives the information in both English and Hausa, as text and audio, alongside relevant clinical guidance that can help her notice anything she might have overlooked. The second face is designed for women at home, who can ask questions in Hausa or English and receive answers in the language they used, with sensitive information such as medication details kept within the health worker version.
Speech recognition and speech synthesis rely on Meta’s Massively Multilingual Speech models, which the team is fine-tuning so the voice women hear sounds natural and familiar. Hausa-speaking team members review the outputs, and the team is evaluating the system with native speakers before it is hosted on the sandbox.
Because network coverage in the area is poor, WOGAFH has designed key features to work offline. Health workers already record a handful of vital signs, such as blood pressure and oxygen levels, on paper forms as patients arrive. Entered into the app, these produce an immediate indication of how urgent a case is, with no internet connection and no shared language required. Questions that need the AI are stored on the phone and sent as soon as a connection returns. The tool runs as an app on any Android phone, since most health workers in this rural setting work from their own devices.
When the team brought paper designs and later higher-fidelity screens to the health centre, nurses responded with real enthusiasm and recorded their feedback, which the team is now reviewing to decide what to integrate before testing. A medical doctor who serves in the local government health office has reviewed the questions the tool asks, and the local government has issued a letter approving the use of the tool by community health workers in its health centres. Once the prototype is complete, WOGAFH will demonstrate it to local officials, agree on which health centre to start with, and present it to the state Commissioner for Health.
The team is completing its prototype for delivery at the end of November 2026, followed by testing with health workers and patients. Looking further ahead, WOGAFH is exploring ways to reach Hausa-speaking women where they already gather and listen, including local Hausa-language radio, short audio programmes shared from phone to phone, and information played in health centre waiting rooms.
The prototype will be presented in the A+ Alliance prototype showcases in early 2027. Follow the A+ Alliance for updates on WOGAFH and the other teams in the cohort.
Photo credits: Fanny Maurel & Digit / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/