Artificial Intelligence-Driven Innovations in Obstetric and Primary Care: Clinical Applications and Future Prospects

Background: AI is already integrated throughout the obstetric and primary care spectrum, ranging from automated ultrasound diagnosis interpretations, foetal monitoring, risk stratification for obstetric complications, to diagnostic decision support in primary care. Objectives: The aim of this review is to identify, summarize and highlight the literature on innovations in obstetric and primary care using artificial intelligence from 2020 to 2025, and to explore clinical applications, reported performance and challenges for use in daily practice. Method: Structured search of Scopus, Web of Science and PubMed-indexed sources was conducted using combinations of the terms “artificial intelligence,” “machine learning,” “obstetrics,” “primary care,” “clinical decision support,” and specific applications (preeclampsia, gestational diabetes and ultrasound) between 2020 and 2025. Findings: AI systems now have the ability of detecting foetal abnormalities and foetal biometry in the hospital setting and are approaching the accuracy of experts in obstetrics, with deep-learning models achieving even higher precision for foetal biometry and anomaly detection on ultrasound when compared with conventional risk-scoring methods. In the obstetrics setting, machine-learning models incorporating maternal traits, biophysical markers, and laboratory values predict preeclampsia and gestational diabetes, with areas under the receiver operating characteristic curve generally between 0.75 and 0.90, consistently outperforming traditional risk-scoring methods. AI-powered clinical decision support tools are being used in primary care to assist with diagnosis, risk stratification and documentation, but widespread adoption of these systems in the real world is more likely to be used for ambient scribing, decision support and workflow automation. In both areas, evidence quality is inconsistent, external evidence is scarce, and integration in clinical practice is inadequate compared to the development of the algorithm. Conclusion: AI is moving beyond the realm of research focus to become a new and growing clinical tool in obstetric and primary care; to achieve the potential of better outcomes for maternal, foetal, and population health, prospective, external evidence is needed and workflow integration, equity and clinician trust must be deliberately addressed.