Spatial agent systems
Practical introductions to how spatial agents use evidence, tools, GIS software, durable context, and verification to complete geospatial work. The evaluation side of this research lives in our geospatial AI agent benchmark registry.
Evidence status: published parts combine current design principles with bounded, practice-led examples. They do not establish a finished Axis Spatial product or autonomous operation.
How to evaluate a spatial agent system
How to assess the task, context, model, tools, runtime, artefacts, verifier, cost, recovery, and limits as one system.
- →Evaluate the complete configuration
- →Inspect artefacts and verifier evidence
- →Record cost, recovery, and limits
Start with Part 1:
What a spatial agent needs
Start with the spatial meaning, evidence, tools, durable context, and verification a system needs before anyone relies on its result.
Geospatial on Databricks
Practical guide for teams running geospatial on Databricks. Volumes traps, Jobs API patterns, and critical mistakes to avoid.
ArcPy to Cloud Migration
Migrating from desktop GIS to cloud-native pipelines. The business case, open-source equivalents, and platform selection.