Floodbase utilizes AI models to assess flood risks. Their US Daily Flood model, integrates diverse data sources such as the National Water Model, USGS stream gauges, NWS rainfall data, and Sentinel-2 satellite imagery. This model generates accurate flood inundation maps at a 250m resolution, hourly since 2019 and every three hours from 1981 to 2019. It significantly improves upon previous models and excels at capturing flash floods, even those difficult to map, such as the Central Texas floods of July 2025. The model's accuracy is validated using NFIP claims, client reports, and media coverage, with AI based large language models (LLMs) assisting in analyzing news for validation.

Floodbase is also developing a new daily global flood model, which similarly leverages AI to combine satellite observations with rainfall, stream gauge data, and hydrological inputs. This ongoing work highlights the growing parametric flood insurance market and the transformative impact of AI in driving innovation in insurance products and scientific research.