Building a better future for female health

Centralizing data to address scalability and innovation

Flo Health, the #1 women’s health app worldwide, is dedicated to helping women understand and take charge of their health through advanced technology and reliable health recommendations powered by data and AI. As Flo Health experienced rapid growth, their previous data infrastructure faced significant hurdles in delivering on their goals. Initially operating with traditional databases and later Trino, scalability issues emerged as data volumes surged. The decentralized data systems and siloed teams emerged as challenges that hindered innovation and the ability to rapidly test and deploy new features. The complexity of managing data across disparate systems also hindered their ability to deliver cohesive user experiences and accurate health predictions.

“Our previous systems couldn’t handle the scale and complexity of our data. Databricks provided the centralized platform we needed to innovate faster,” Roman Bugaev, CTO at Flo Health, explained.

Furthermore, the desire to democratize access to data for self-serve analytics was stifled by business data users who were limited to SQL queries, creating an over-reliance on the engineering team to build reliable and secure data pipelines for downstream analytics. Transitioning to the Databricks Data Intelligence Platform allowed Flo Health to not only unify their data and analytics operations but also simplify secured access to data for the entire organization. By adopting a modern lakehouse architecture, they centralized their data, eliminating silos and enabling domain-specific analytics teams to collaborate more effectively.

Boosting innovation through a unified data architecture

With the Databricks Data Intelligence Platform, Flo Health has revolutionized how they manage and utilize data. The platform supports their ability to run 150–200 experiments simultaneously and 400 per quarter, driving insights into user preferences and feature performance. Their AI models now power advanced health predictions, offering users actionable insights for conditions like irregular cycles or dietary needs during pregnancy.

“The ability to run experiments at this scale has fundamentally changed our approach to product development,” Dmitry Zolotukhin, VP of Analytics at Flo Health, noted.

With Delta Lake, the company has established a medallion architecture to help Flo Health provide the right data for both analytics/business intelligence (BI) and machine learning (ML) use cases. This includes seamless support for managing and analyzing their ML experiments, with Databricks SQL powering analytics and BI workflows that are visualized through business dashboards in Looker and technical dashboards within Databricks. To meet stringent compliance regulations around women’s health privacy, Flo Health uses Unity Catalog to enforce fine-grained access controls and robust data governance throughout their data workflows. This approach helps avoid any issues with personally identifiable information (PII) but also serves as a building block for award-winning features like Anonymous Mode, which anonymizes sensitive customer data to further enhance privacy.

With data flowing through pipelines reliably and securely, Flo Health leverages MLflow to streamline machine learning operations — allowing them to efficiently manage the ML lifecycle, including model versioning and deployment. While their migration to serverless compute is ongoing, this feature has already reduced operational overhead and improved scalability. The primary motivators for this transition included faster performance at scale and enhanced data security. Flo Health has also been piloting AI/BI Genie to democratize data insights through natural language prompts, enabling business users to ask questions via natural language like “What is our MAU for last month?” and receive actionable answers instantly.

Additionally, Flo Health has been fine-tuning large language models (LLMs) for their Health Assistant chatbot using Mosaic AI as the backbone of an agentic AI system, accelerating the development of personalized and accurate user interactions. Lastly, they are an early adopter of Unity Catalog Metrics to further enhance their data-driven insights and the reliability of their reporting capabilities by ensuring business intelligence metrics always line up.

Overall, data democratization — a key priority — has seen a significant boost, with 46% of Flo Health employees, including engineers, data scientists and business analysts, actively engaging with Databricks monthly to make informed decisions using Looker dashboards powered by the platform. This collaboration extends beyond operational improvements, as Flo Health and Databricks are also partnering on research projects aimed at advancing data-driven women’s health initiatives, reinforcing their shared commitment to leveraging technology for societal benefit.

Central to this effort is the use of Databricks Assistant, which has streamlined SQL query debugging, enabling teams to be more efficient and productive. As part of a broader initiative called Copilot, Flo Health is empowering product managers and analysts to work seamlessly with SQL. With Databricks Assistant as the foundation, this feature ensures everyone can access and leverage data insights with ease. Already rolled out to the entire organization, it exemplifies Flo Health’s commitment to making data-driven decision-making accessible to all stakeholders.

Achieving measurable success in data-driven health

The transformation has delivered measurable results. Since adopting Databricks, Flo Health has increased internal adoption, with their monthly active users (MAU) jumping by 45% and weekly active users (WAU) by 57%, underscoring the platform’s ease and powerful functionality. Experimentation throughput and feature releases have skyrocketed, with over 1,293 features and 5,617 configurations now managed in their Experiment Service. This capability enables Flo Health to maintain their position as a market leader while offering unparalleled value to their users.

“Databricks enables us to innovate without limits. The scalability and security have been game changers for our team,” Roman shared. Operational efficiencies have also improved.

Flo Health’s journey with Databricks is far from over. The company is piloting cutting-edge applications like generative AI–powered chatbots to further personalize user experiences. These initiatives promise to deepen user engagement by offering tailored advice for scenarios such as irregular cycles or pregnancy nutrition.

The team is also exploring Databricks Apps to address use cases requiring inputs from business stakeholders alongside advanced model outputs. For example, marketing budget forecasting and optimization — powered by Python models — requires a level of complexity that traditional BI tools like Looker cannot support. With Databricks Apps, Flo Health aims to integrate these workflows seamlessly into their ecosystem, unlocking new opportunities for data-driven decision-making and strategic planning.

“We see endless possibilities with AI and Databricks. It’s about creating smarter, more intuitive experiences for our users,” Dmitry added. With Databricks as their trusted data partner, Flo Health is not just reshaping women’s health — they’re setting the gold standard for what’s possible when innovation meets purpose-driven data strategy.