Zimbabwe Left Its Pregnant Women Alone, and a Chatbot Filled the Gap

Rutendo, 31, from Rusape, went days without feeling her baby move in her final trimester. She did what millions of women in this country now do when fear strikes at night. She asked a chatbot. It told her to see a doctor immediately. She ran to her local clinic, the one where she was already registered for antenatal care, and it was too late. The foetus had become tangled in the umbilical cord and had died.

Let that sink in. A grieving mother had to be saved by an app, and it still was not enough. Ask yourself why a woman in one of our cities had nowhere else to turn at the hour she needed a trained human being.

The answer is not technology. The answer is a health system that has been hollowed out by neglect, underfunding and the exodus of its own people. Zimbabwe had just over 22 health workers per 10,000 people in 2022, according to a WHO and Ministry of Health analysis. The WHO's recommended minimum is 45 per 10,000. That is less than half. Every midwife who is missing from a rural ward is a woman who labours without someone who knows what to look for. Every nurse who has emigrated is a queue that grows longer by the day.

Now look at what DawaMom, an AI chatbot built by DawaHealth, a Zambia-based startup, is actually doing. It runs on WhatsApp, the app that Southern Africans already pay for in small data bundles, because the open internet is too expensive for most families. Users can ask about nutrition, back pain, frequent urination and foetal movement, and the service connects them to licensed health workers. The product is free to ask questions. Personalised monitoring and home visits cost money, which tells you who the business model is really built for.

The design choices are worth examining, because they reveal the difference between a serious health tool and a viral gimmick. DawaMom does not scrape the internet and guess. It uses retrieval-augmented generation, meaning it answers only from a verified database. According to the company, that database was compiled from Health Ministry guidelines, World Health Organization recommendations, medical school curricula and hospital protocols on women's health. More than 25 health workers translated and validated the content into Shona, Ndebele and Tonga. The team also built an English-Shona code-switching dataset with more than 30 Zimbabwean women, so the bot understands how people actually talk, mixing languages mid-sentence, and replies in a voice that sounds familiar.

That is what respect looks like in practice. Zimbabwean women are not a demographic to be "reached." They are mothers who speak in their own languages and deserve answers in them.

Now the honest part, because Revvmark does not ask you to swallow claims whole. The figures we have come from the company itself. DawaMom has supported more than 20,000 users and targets 50,000 by the end of 2026. Those numbers are not independently audited. The stories of Samantha Murungweni, who delivered a healthy girl last month, and Rutendo, whose outcome was tragic, are powerful testimony. They are not a clinical trial. We still do not know how often the triage flags are wrong, how many high-risk women the system misses, or what happens to the health data these women share over WhatsApp. Zimbabwe has a data protection law. Companies handling the most intimate information a woman will ever share must be held to it, publicly and without exception.

Here is the uncomfortable question. Why are we depending on a Zambian startup, a WhatsApp subscription model and the goodwill of donors to deliver what the state is constitutionally obliged to provide? Zimbabwe's citizens are entitled to competent antenatal care. That obligation does not disappear because a private company has built a clever tool to partly fill the void.

The solutions are not mysterious. Africa already has proof that they work. Rwanda built one of the continent's most respected community health models, training and deploying community health workers who bring care to villages that the nearest clinic cannot reach. That model is worth studying, even as we condemn Kigali's military adventures in eastern Congo, because a government can be effective at home and destructive abroad. Botswana built its health and governance credibility on institutional discipline and consistent public investment. Mauritius shows what happens when a country treats competent public services as a national priority rather than an afterthought. None of these success stories were built by accident. They were built by governments that funded health workers, published their data, and answered to their citizens.

Zimbabwe can do the same. The government should fund and fill vacant midwife and nurse posts, with transparent public tracking of staffing levels in every district. It should set clear regulatory standards for AI health tools, requiring independent clinical validation, published error rates and the right of users to delete their data. Ministry guidelines should be made public so that any digital service can be checked against them. Professional bodies like ZiMAIN should be given a formal role in auditing these tools, not just cheering them on from the sidelines. Dakwa is right that AI should support clinicians rather than replace them. Support is only meaningful if the clinicians exist.

Civil society and the private sector have roles to play too, but they must stay honest about their limits. Startups should publish their validation methods, submit to independent review and resist the temptation to turn a mother's fear into a recurring subscription. Donors should fund the human workforce first and the apps second, because no algorithm can deliver a caesarean section or resuscitate a newborn.

So here is what we demand. Demand that the Ministry of Health publish current staffing data for every district and a dated plan to close the gap with the WHO benchmark. Demand that any AI health service operating in Zimbabwe disclose its validation results and data practices in plain language. Demand that every pregnant woman, in Rusape, in Mutare, in Harare's suburbs, can reach a qualified human being within a reasonable time, without paying for a data bundle or a subscription.

Rutendo's baby should not have been lost in the space between a warning and a clinic door. Samantha's healthy daughter should not be the exception that makes the system look like it is working. Every woman who opens WhatsApp at two in the morning, terrified, deserves better than a chatbot and a prayer. Our leaders are responsible for that standard, and we are responsible for refusing to accept anything less.

Based on reporting by RFI