In Kenya, where access to medical specialists and internet connectivity can be severely constrained, an artificial intelligence platform operating over basic text messaging is helping expectant mothers navigate pregnancy and secure vital medical attention. The system, known as PROMPTS, was developed by Jacaranda Health to triage patient inquiries, deliver accurate healthcare guidance, and rapidly connect women in distress with trained clinical staff.
The widely used digital tool originated entirely by happenstance. Years ago, a network of maternal health clinics operated by Jacaranda Health implemented an automated Short Message Service (SMS) program designed solely to broadcast clinic appointment reminders to pregnant patients. Rather than treating the notifications as one-way alerts, however, recipients quickly began using the number to write back with pressing health concerns.
The incoming messages covered a wide spectrum of maternal health topics. Expectant mothers frequently sought advice on day-to-day nutritional safety, asking whether foods such as avocados were permissible to eat during gestation. Others reached out seeking clarification about physical symptoms they were experiencing, inquiring whether bodily swelling, severe headaches, or other discomforts fell within the normal range of pregnancy.
In the initial stages of the project, the volume of incoming texts was modest. Jay Patel, the director of technology at Jacaranda Health, noted that when he first joined the organization, the platform handled fewer than 100 inquiries on a daily basis. As awareness spread among clinic attendees, however, the message traffic expanded exponentially, eventually growing into a massive stream of roughly 15,000 questions each day.
To manage the early influx of communication, Jacaranda Health initially hired clinical nurses tasked with reviewing and answering every message individually. Yet, as thousands of new inquiries poured in daily, manual message processing by human staff became unsustainable, prompting the organization to explore technological solutions capable of organizing and prioritizing the high volume of texts.
Beginning in 2020, software developers began constructing an artificial intelligence framework designed to comprehend the nuances of incoming patient messages. The resulting PROMPTS program analyzes the text, categorizes the subject matter, prioritizes queries based on clinical severity, and determines whether an individual user might require immediate, in-person clinical care.
Today, the PROMPTS system serves a wide footprint across Kenya, operating in public healthcare facilities across 24 counties. By integrating directly into public sector clinics, the platform provides continuous support to mothers who rely on government healthcare infrastructure for their prenatal and postnatal checkups.
The reliance on standard SMS infrastructure is one of the platform’s most significant design strengths. In many regions of the country, mothers do not own modern smartphones, lack consistent cellular data connections, or cannot easily access web search engines. Under these constraints, a simple text to the PROMPTS platform often represents the quickest and most dependable method to receive accurate medical advice.
The ability to receive rapid medical guidance also solves a crucial logistical dilemma for families residing far from healthcare facilities. In many instances, the automated service helps patients evaluate their symptoms and determine whether undertaking an arduous and expensive journey to a regional hospital is clinically necessary.
Patel emphasized the vital nature of this connectivity for rural populations facing sudden obstetric complications. If a pregnant woman in a remote area experiences bleeding in the middle of the night, such as at 2 a.m., she rarely has the ability to search the web for an answer. Because the PROMPTS service operates entirely over SMS and is free to users, it remains accessible to anyone equipped with even an entry-level mobile handset.
While the artificial intelligence model handles the majority of routine inquiries, it includes safeguards to detect critical medical emergencies. Jacaranda Health estimates that approximately 7% of all incoming text messages are flagged by the system as potentially urgent clinical cases.
Whenever a message is categorized as an emergency, it immediately bypasses all automated AI response mechanisms. The system routes the message directly to an emergency queue at a dedicated help desk staffed by human healthcare professionals.
Once an urgent query enters the queue, nurses can review the user's past medical question history and immediately initiate an outgoing telephone call to the patient. This direct human intervention ensures that high-risk cases receive immediate assessment without relying on synthetic responses.
The necessity for rapid clinical triage is underscored by the severe maternal health challenges facing the nation. In Kenya, approximately 6,000 women lose their lives every year due to complications related to pregnancy and childbirth, highlighting an urgent need for interventions that detect danger signs early.
By transforming a routine appointment reminder tool into an intelligent triage and counseling network, the PROMPTS platform demonstrates how low-bandwidth technology can bridge healthcare access gaps and provide a critical safety net for thousands of mothers across the country.
(0)