Kermit GeoAI at UIS 2026 in Tönning
Analyzing geodata, integrating standards, keeping data secure - simplifying GIS analyses with GeoAI
Anyone who works with geodata daily knows the pattern. A specialist question is clear: Which protected areas are located in the corridor? What distances must be maintained? Where do conflicts arise with existing habitat areas? The data is available, the GIS is open. And yet, hours pass before a reliable answer is available. Not because tools are missing, but because every analysis still involves many manual steps: merging data, combining layers, documenting results, and finding the relevant standard that legally substantiates the result.
At the 33rd Workshop of the GI Specialist Group Environmental Information Systems in Tönning, hosted by the State Agency for Coastal Protection, National Park Administration and Marine Protection Schleswig-Holstein, MovingLayers demonstrated how Kermit GeoAI makes these steps executable via simple text input: fully locally, with complete documentation, without data leaving your own infrastructure.
Coastal protection, infrastructure, nature conservation: the same question in different contexts
MovingLayers works with network operators, pipeline companies, and infrastructure planners; companies that carry out impact assessments, restriction checks, and route evaluations daily. Kermit was developed precisely for this daily work: load protected areas, calculate buffers, analyze overlaps, output results with sources, as a reproducible and auditable process.
The workshop in Tönning showed that the UIS community experiences the same daily routine, in a different professional guise. Instead of pipeline routes, it's about dike sections; instead of concession areas, it's about Natura 2000 sites; instead of network expansion planning, it's about habitat monitoring and species protection. The basic question is the same: How can you get from a specialist question to a spatial analysis quickly, securely, and transparently? And it affects specialists with and without a GIS background equally.
The example from our presentation, which dike sections intersect designated habitat areas in the foreland, was deliberately tailored to the host. The fact that the modernization of the Schleswig-Holstein nature conservation database was presented on the second day of the conference, including new modules for invasive species, underscored this connection. Such data sets are available. What Kermit contributes is the direct, secure path from the specialist question to the analysis: without detours via a GIS specialist, without manual intermediate steps.
Standards as part of the analysis, not as rework
In planning and administrative processes, the actual result of a spatial analysis is not the map, but the legally sound statement behind it. Which standard justifies the distance? Which paragraph of the Federal Nature Conservation Act protects this area? Which provision of the development plan is affected? These questions are usually answered downstream today: after the GIS analysis, in a separate research step.
MovingLayers is actively working to integrate this step into Kermit. The goal is an AI-optimized integration of standards, initially from building law, and prospectively from other planning and environmental law areas, directly into the analysis process. Kermit will then know the relevant regulations, use them in the analysis, and output the result with the appropriate standard reference. An impact analysis not only delivers the affected areas but also the legal planning basis on which the impact is based.
For infrastructure planners and network operators, this closes a gap that costs considerable time in daily work. For the UIS community, where standards such as the Habitats Directive, the Federal Nature Conservation Act, or state-specific nature conservation laws form the professional basis of every evaluation, it opens up a new quality of analysis documentation: not only what was found, but why it is relevant.
No shortage of data
The workshop showed how rich the data available in environmental informatics already is. The Federal Agency for Nature Conservation and con terra presented the status of data management for marine habitats and benthos; Kiel University showed how satellite data can quantify seagrass biomass. This data is there: maintained, standardized, increasingly openly available. What is still often missing is the fast, secure path from the data set to the legally classified specialist analysis, without every evaluation ending up on the desk of a GIS specialist.
The Federal Environment Agency and the Federal Agency for Cartography and Geodesy posed precisely this question in their presentation on the operationalization of GEO-AI in federal administration: How can GeoAI capabilities be integrated into the everyday work of authorities without compromising data security and traceability? Kermit is a concrete answer to this; not a conceptual one, but one that can be integrated into QGIS today.
Secure, local, auditable: because KRITIS and the EU AI Act are not options
In infrastructure operations and public authorities, the rule is: Anyone who uses GeoAI must be able to explain how a result came about and ensure that sensitive geodata does not leave their own environment. Kermit is built for these requirements. Processing takes place completely locally, on-premise or in a controlled private cloud environment. Every analysis automatically generates a complete chain of evidence: layers used, parameters, standard references; auditable and verifiable.
The standard integration we are working on adds another dimension to this chain of evidence. The analysis then knows not only what it has done, but why the result is professionally and legally sound. In approval procedures, environmental reviews, and official planning documentation, this is not a convenience feature, but a basic prerequisite.
Two contributions, one claim
In addition to the Kermit presentation, MovingLayers was represented with a second contribution. Julia Hacker-Zagler, together with the University of Salzburg, presented Virtual Shepherd: an IoT-based assistance system for automatic pasture monitoring in the alpine region. Different domain, same claim. Spatial data should reach where decisions are made; without technical hurdles between question and answer.
A frog doesn't choose between water and land. It needs both – and it moves freely between them. Not as a compromise. As nature. The same applies to working with geodata and standards. Analysis or classification. GIS tool or specialist standard. Technical evaluation or reliable justification. Kermit is not a tool for one of these sides – it is a leap ahead: it plans across domains before the first step is executed.
If you want to know how Kermit GeoAI works in your own GIS environment – whether in environmental, coastal protection, or infrastructure contexts – you'll find a concrete introduction in the whitepaper.
In the whitepaper, we describe the architecture, use cases, and prerequisites of Kermit GeoAI – as a basis for an initial conversation or for your own assessment.
→ Download Whitepaper→ Request initial consultation
Recommended reading
