GeoAI Agency Primitives

2026-04-02

AI Summary

While existing GeoAI models possess basic vision-language capabilities, they fall short in effectively supporting GIS practitioners in real-world tasks such as vector layer editing, raster processing, and cartographic production. To bridge this gap, this work proposes an agent capability layer tailored for geospatial intelligent assistants, systematically defining for the first time nine core, implementable, testable, and comparable primitives—spanning navigation, perception, georeferenced memory, and dual modeling—to enable human-in-the-loop iterative collaboration. By integrating large language models, GIS workflows, and human–AI coordination mechanisms, the framework not only narrows the divide between foundational model capacities and practical GIS demands but also establishes a human productivity–centered evaluation benchmark, offering a viable technical pathway and assessment standard for deployable GeoAI assistants.

Abstract

We present ongoing research on agency primitives for GeoAI assistants -- core capabilities that connect Foundation models to the artifact-centric, human-in-the-loop workflows where GIS practitioners actually work. Despite advances in satellite image captioning, visual question answering, and promptable segmentation, these capabilities have not translated into productivity gains for practitioners who spend most of their time producing vector layers, raster maps, and cartographic products. The gap is not model capability alone but the absence of an agency layer that supports iterative collaboration. We propose a vocabulary of $9$ primitives for such a layer -- including navigation, perception, geo-referenced memory, and dual modeling -- along with a benchmark that measures human productivity. Our goal is a vocabulary that makes agentic assistance in GIS implementable, testable, and comparable.

Problem

• GeoAI
• agency primitives
• human-in-the-loop
• GIS workflows
• productivity gap

Innovation

• GeoAI
• agency primitives
• human-in-the-loop
• geospatial workflows
• foundation models