Ask the Earth a Question: Planet Turns Satellite Imagery Into Chat

Planet, a satellite imaging company, has released an app that lets you search the Earth in ordinary language. Instead of learning a geographic information system, a user types something like “find new construction in this area” or “track changes in this forest”, and the model locates the place, pulls the relevant imagery and shows the result.

The product is free at ai.planet.com with a Planet Insights account, with more advanced features planned. It draws on more than a decade of the company's global imagery.

There are three capabilities. The first is search: understanding what a visual feature looks like and finding it across a wide area. The second is change: going back through ten years of archive, identifying when something changed, and producing a side-by-side comparison or a time-lapse. The third is analysis: handing a more complex task to an agent, with uploads of images or CSV files, scheduled trend tracking and recurring reports.

The design decision worth noting is the insistence that every answer be traceable. Planet says all AI conclusions rest on imagery the user can open and inspect, a deliberate hedge against the failure mode that has undermined similar tools — a confident sentence with nothing behind it.

Early testers have used it for conservation, land management and environmental compliance. A conservation team can watch a habitat, spot illegal encroachment or fire risk, and check a deforestation alert before sending rangers out. A compliance team can filter, in minutes, the industrial sites that need inspection because they sit near a waterway — work that used to take days of manual GIS analysis.

Planet files this under a strategy it calls “planetary intelligence”: pairing daily satellite observation with AI so that changes on the ground are understood close to real time.

The broader shift is from remote sensing as a specialist skill to something closer to a public tool. That depends on three things holding up — coverage, refresh rate, and whether results can be explained — and if they do, the obvious customers are agriculture, disaster response, city government, insurance and infrastructure.