Building trusted agentic AI for maternal health at scale

Maternal health outcomes in the United States remain deeply inequitable, with underserved communities—especially Black and Indigenous women—disproportionately affected by pregnancy-related complications and loss. More than 80% of pregnancy-related deaths are preventable.¹ The CDC’s definition of preventability—deaths that could be avoided with reasonable changes to patient, provider, facility, system and community factors¹—helped sharpen MyLÚA Health’s founding insight.

Improving outcomes requires trusted support that reaches families beyond the clinic. That conviction sparked LÚA, an AI-powered maternal-wellness platform designed to deliver evidence-based, culturally attuned guidance throughout pregnancy and the first year postpartum, including the critical moments between appointments. “Every mother deserves support that understands her—not just her chart. LÚA was built to meet families where they are, with guidance that reflects their culture, their language, and their lived experience,” said J’Vanay Santos-Fabian, MBA, CEO and Co-founder of MyLÚA Health.

To make that vision practical at scale, MyLÚA Health needed an enterprise-ready platform that could provide consistent answers, personalize support with empathy and context, protect sensitive health information, and connect users to next steps such as check-ins, reminders and local resources.

MyLÚA partnered with IBM to design, test and deploy its multi-agent maternal-wellness platform underpinned by IBM watsonx.ai®, IBM watsonx Orchestrate®, IBM Cloud® and IBM Cloud Code Engine. The goal: deliver empathetic, context-aware support that can meet users where they are, understand needs and take action—safely and at scale.

IBM watsonx.ai provides a governed model environment with multilingual capabilities and responsible AI practices suited for healthcare use. IBM watsonx Orchestrate coordinates specialized agents across emotional, physical and social domains, turning user intent into completed actions such as check-ins, reminders, resource activation and summaries. Deployed on IBM Cloud using IBM Cloud Code Engine (with FastAPI, Postgres and Redis), the architecture supports secure scaling for regulated workloads. After extensive user acceptance testing, the agents were deployed and the platform is available in the Apple App Store and Google Play Store.

In early validation, MyLÚA’s patent-pending multimodal AI model demonstrated over 90% accuracy in identifying postpartum depression risk during the first trimester—months before traditional screening methods. Trust signals were similarly encouraging: 79% of mothers said they felt comfortable sharing sensitive health information through the platform. Engagement indicators were similarly encouraging: 48% of participants completed proactive wellness activities and 44% accessed trimester-specific education.

For care teams and maternal-health organizations, early pilot testing showed the potential for earlier signal generation—64% of mothers completed health risk assessments—while also reducing administrative burden through structured summaries of member interactions that help doulas and care coordinators spend more time on high-touch support. The platform’s privacy-preserving design further supports healthcare-grade governance and scalability because personally identifiable information and protected health information are not sent to large language models.

Today, MyLÚA’s multi-agent platform is live, expanding access to maternal-wellness support with multilingual, culturally attuned guidance for pregnancy and the first year postpartum. LÚA also goes beyond answering questions by completing support actions such as automated check-ins, reminders and resource activation, helping families get timely help between appointments. Michael Conward, Ph.D., CTO and Co‑founder of MyLÚA Health said, “We created LÚA to ensure no woman has to navigate pregnancy or postpartum care alone. Technology allows us to deliver proactive, compassionate support at scale—support that reflects needs of birthing parents, mothers, and pregnancy and postpartum support recipients.”

For care teams and maternal-health organizations, the platform turns conversations into actionable insights and summaries that can reduce manual effort and support larger caseloads. Its modular design can scale with demand and integrate into provider, payer, employer and community-based programs.

By uniting human‑centered design with responsible AI for good at its core, MyLÚA Health and IBM are helping close gaps in maternal care, strengthen trust, and broaden access to high‑quality support—creating a more equitable and compassionate future for women and families.

MyLÚA Health is an enterprise agentic AI platform built on IBM watsonx that delivers personalized, multilingual maternal wellness support. Its patent-pending multimodal AI framework identifies early risk patterns from diverse data inputs, transforming emotional, physical, and social signals into actionable insights for mothers, care teams, and health organizations.