Recent advances in geospatial foundation models, such as Google’s Population Dynamics Foundation Model (PDFM), are opening new possibilities for understanding human behavior in cities. These embeddings encode complex patterns of population movement, activity, and interaction, yet translating them into actionable insights remains a major challenge. This talk will introduce the integration of Google’s Population Dynamics Foundation Model (PDFM) embeddings into the CARTO ecosystem. I’ll discuss what these embeddings capture about urban behavior, the common barriers to adoption we’ve observed among users, and our approach to lowering those barriers through purpose-built tooling. I’ll highlight our new Workflows extensions, which enable clustering, similarity analysis, visualization, anomaly detection, and traditional ML directly on embeddings—bridging the gap between foundation models and everyday spatial analysis. A live demo will illustrate how these capabilities can turn complex behavioral data into clear, actionable insights.