Equity in Women’s Health

Marie.AI

Equity in Women’s Health

Marie.AI organizes women’s health data with AI, improves gynecological care, and informs evidence-based public policies.

Ribeirão Preto, SP, Brasil

BRA

Nonprofit

Ovarian cancer is one of the deadliest conditions in women’s health: in 2022, 324,000 women were diagnosed worldwide, and 70% died due to late detection. This reflects deeper systemic neglect — driven by fragmented health data, outdated protocols, and an individualized, reductionist view of the female body that ignores the complexity of women’s lives. Benign conditions such as polycystic ovary syndrome (PCOS), chronic pelvic pain, and fibroids are still treated with protocols from the 1980s and 1990s, often using hormonal contraceptives or antispasmodics. In the case of ovarian cancer, innovation and research remain scarce, especially compared to areas like breast cancer. Marie.AI addresses this critical gap by using artificial intelligence to organize real clinical data, identify risks through evidence-based models, and generate personalized care recommendations. It strengthens the doctor–patient relationship through what we call a "return to eye-to-eye care", while also enabling individualized treatment planning. More than a diagnostic tool, Marie.AI lays the foundation for dynamic clinical protocols, improved early detection, and smarter, data-driven public policies that reflect the true needs of women — especially those historically excluded from innovation in healthcare.

The platform analyzes underused data such as lab results, symptoms, and medical history — already stored in health systems — and combines this information with non-radiation imaging (like ultrasound) to identify gynecological diseases early, accurately, and safely. Marie.AI was developed over 10 years of research, integrating knowledge from neuropsychology, neuroanatomy, and electrophysiological studies of hippocampal neurons, fusiform cells, cones, and rods. These foundations helped us design algorithms that improve data concordance between image analysis, hemogram patterns, and clinical history, enhancing the system’s precision and relevance in real-world healthcare. Instead of requiring expensive equipment or specialist teams, Marie.AI works via mobile phone, tablet, or web app. Health professionals can take a photo, upload an exam, or connect to an electronic medical record. The AI processes the data and returns insights in under 30 seconds. Marie.AI uses a Variational Autoencoder (VAE) enhanced with Minimum Description Length (MDL) regularization, and applies the principle of parsimony to differentiate benign, rare, and malignant conditions. It supports clinical decision-making and generates structured data for smarter public health policy.

Marie.AI directly serves cisgender and transgender women, and people with uteruses, with a special focus on populations historically neglected by the healthcare system — including Black, Brown, low-income, riverside, Indigenous, and underserved urban communities. These populations often face late or missing diagnoses, overuse of radiation-based exams, generic treatments, and clinical neglect. Gynecological symptoms are frequently misunderstood or minimized due to the lack of structured data and access to trained specialists. Marie.AI addresses these gaps by transforming basic clinical data (like hemograms and medical history) and radiation-free imaging (like ultrasound) into intelligent, personalized diagnoses in under 30 seconds, using AI developed through 10 years of scientific research. The platform runs on mobile phones, tablets, or web apps, requiring no specialized equipment. This allows Marie.AI to bring advanced diagnostic support to remote and low-resource areas, saving lives through early detection, reducing historical health disparities, and strengthening women's autonomy in their own care — regardless of geography or socioeconomic status.