Master ArcGIS GeoAI Toolbox with our step-by-step tutorial, detailed feature walkthrough, and expert tips.
Explore the key features that make ArcGIS GeoAI Toolbox powerful for automation & workflows workflows.
The GeoAI Toolbox is a geoprocessing toolbox inside ArcGIS Pro that trains and runs AI models on geospatial and tabular data. It contains over 30 tools organized into four toolsets: Feature and Tabular Analysis, Imagery AI, Text Analysis, and Time Series AI. These toolsets cover classification, regression, object detection, pixel classification, natural language processing, entity extraction, and forecasting on space-time cubes. The toolbox supports more than 50 deep learning architectures including U-Net, Mask R-CNN, FasterRCNN, DeepLabV3, and transformer-based models, using both classical machine learning and modern deep learning techniques integrated directly with GIS layers.
The toolbox itself is included with ArcGIS Pro, so the cost is essentially the cost of an ArcGIS Pro license. A Basic named user subscription starts at approximately $700 per year. Standard licenses run approximately $1,400 per year, and Advanced licenses approximately $2,200 per year. The Image Analyst extension, required for most deep learning imagery tools, adds approximately $1,400 per year. Enterprise and academic institutions often have site licenses that include access at no additional per-seat cost. All prices are Esri list prices and may vary by region and agreement.
Yes. The documentation explicitly notes that all tools in the GeoAI toolbox require the installation of deep learning framework libraries such as PyTorch, TensorFlow, and fastai. Esri provides a dedicated Deep Learning Libraries Installer that matches these dependencies to your ArcGIS Pro version. Without this installer, most tools in the Imagery AI and Text Analysis toolsets will fail to run. GPU-capable hardware is also strongly recommended for training deep learning models in any reasonable time.
You can absolutely use pretrained models. ArcGIS Living Atlas provides over 100 pretrained deep learning models for tasks like building footprint extraction, road extraction, land cover classification, and object detection across various sensor types. The Text Analysis toolset specifically supports fine-tuning pretrained NLP models, and you can apply imagery models directly to your data, fine-tune them on your own labeled samples, or combine them with models built in the ArcGIS API for Python arcgis.learn module. This dramatically reduces the labeled data required for production-ready results.
The GeoAI Toolbox trades openness for integration. Open-source stacks like QGIS with scikit-learn, rasterio, and PyTorch are free and flexible but require manual plumbing between spatial data formats, ML libraries, and visualization tools. GeoAI Toolbox handles all of that inside ArcGIS Pro, with native support for feature classes, rasters, and space-time cubes, and output layers that drop straight into maps. The tradeoff is the licensing cost (starting at ~$700/year for Basic, plus extensions) and lock-in to the Esri ecosystem.
Now that you know how to use ArcGIS GeoAI Toolbox, it's time to put this knowledge into practice.
Sign up and follow the tutorial steps
Check pros, cons, and user feedback
See how it stacks against alternatives
Follow our tutorial and master this powerful automation & workflows tool in minutes.
Tutorial updated March 2026