Geographers

ESRI's 2023-2025 product roadmap has embedded generative AI throughout ArcGIS — ArcGIS Copilot allows users to execute spatial analysis through natural language, ArcGIS AutoML builds predictive models with no coding required, and AI-generated map narratives automatically describe analytical outputs. Google Earth Engine now supports AI-assisted script generation, and Microsoft Planetary Computer's integration with Azure OpenAI allows LLM-orchestrated geospatial analysis pipelines. Smaller platforms like Felt, Carto, and Foursquare Studio are building AI-first interfaces from the ground up that assume non-specialist users.

NASA and IBM released Prithvi, a 100-million parameter geospatial foundation model trained on six bands of Harmonized Landsat Sentinel-2 imagery, which achieves state-of-the-art performance on flood mapping, wildfire scar detection, and crop segmentation with minimal fine-tuning. The Clay foundation model processes multi-modal satellite data (optical, SAR, multispectral) with zero-shot generalization across geographies. Google's DynamicWorld delivers continuous 10m global land cover classification updated weekly. ESA's Copernicus Emergency Management Service automated rapid mapping achieves activation-to-product times measured in hours for disaster events that previously required days of analyst work.