The hype around AI feels all-encompassing. It's impossible to escape in the news, in social media, in everyday conversation, and especially anywhere in the tech industry, of which geospatial is a part.
But instead of just going on about how it's going to revolutionize everything, we partnered with Geoawesome) to create a research report examining its use cases in Earth observation through a more critical lens. You can download the report for free via the button below.
The report is titled AI in Earth Observation: What's Real and What's Next. To build it, we used the Gartner Hype Cycle, a framework that describes how technologies evolve over time, from early buzz to practical use. Here are the five stages:
-
Innovation trigger: A breakthrough, concept, or prototype generates excitement. At this stage, we see early demos and media interest, but practical implementation is minimal.
-
Peak of inflated expectations: A frenzy of hype builds up. Expectations soar as early successes are overgeneralized, often leading to lofty promises that the tech will solve everything. Many initiatives at this stage are not yet proven at scale.
-
Trough of disillusionment: Reality sets in. Initial enthusiasm collapses when pilots fail to deliver as hoped. Some projects are abandoned or pivoted. The technology’s limitations become apparent, leading to a more sober perspective.
-
Slope of enlightenment: The technology claws its way out of the trough. Practical learnings emerge; developers and users better understand where it can actually add value. Improved iterations appear, and adoption grows slowly but steadily, often
in niche areas or pilot projects that demonstrate real impact.
- Plateau of productivity: The innovation becomes mainstream. It’s widely adopted and proven; the benefits are clear and validated. Hype is replaced by real-world productivity.
At this stage, the technology is no longer“new;” it’s a standard tool delivering reliable results.
Now, it's important to note something here: just because a use case is in stage 2 or stage 3 does not mean it's failed. This is a natural progression that many exciting technologies go through. It just means that tech's not yet ready for widespread adoption. Many of the AI use cases in the report are in stage 2 or 3 currently, but still show plenty of promise.
We looked at foundation models, autonomous AI agents, real-time onboard AI for satellites, agentic EO workflows, and spatiotemporal data cubes for machine learning, along with many other use cases.
Want to find out more? We won't spoil it in this post. Download your copy for free here. Or, if you have questions, feel free to reach out to us here. We'd be happy to get you some answers.
