Background <p>Despite Sri Lanka’s long-standing success in reducing maternal and neonatal mortality, progress has stagnated in recent decades. Conventional surveillance systems lack granular, real-time data to predict and prevent adverse outcomes. Artificial Intelligence (AI) has transformative potential to improve risk prediction, early detection, and proactive management of maternal and neonatal complications.</p> Methods <p>This prospective birth cohort study will recruit 2,000 pregnant women in different stages of gestation across all 54 divisional health areas (DHA) in the Western Province, Sri Lanka. Data will be collected using an Electronic Health Record (EHR) System named ‘Maathru’, together with a companion mobile application designed for pregnant women. Data collection will span from the antenatal period through six weeks postpartum. The mobile app captures patient reported data such as information on contraction frequency, general health status, danger signs, heavy physical activity, exposure to violence, sexual activity, and substance use directly from the pregnant mothers.</p> <p>The data gathered from the EHR will be used in the training of AI algorithms to predict key maternal and neonatal outcomes, including postpartum haemorrhage, preeclampsia, intrauterine growth restriction, need for labour induction, gestational diabetes mellitus, caesarean section, postpartum depression, preterm birth, low birth weight, Neonatal Intensive Care Unit (NICU) admission, and feeding difficulties. An iterative process will be adopted in training the algorithm. The AI models were selected based on prior literature and evidence from related studies. The design of the EHR will facilitate integration of the trained AI models, which will then contribute to clinical decision making process supporting the field staff and alerting system for health professionals within the Maathru system and the mobile app respectively.</p> Discussion <p>The study seeks to demonstrate feasibility and validity of AI-powered predictive analytics within routine field-based maternal and child health (MCH) workflows. The mobile application enhance the communication, and monitor maternal well-being. It aims to create a scalable platform for national-level implementation, generate disaggregated data for maternal and neonatal health, and strengthen evidence-based decision making to achieve Sustainable Development Goals (SDGs) related to maternal and neonatal mortality reduction.</p> Trial registration <p>Not applicable (observational study).</p>