0
0
0
17 March 2026The development of artificial intelligence is significantly changing how IT systems operate and the relationship between humans and computers. Naturally, this raises the question: what does this mean for Usemaps?
There are several aspects to consider.
🔵 Will the development of AI cause people to stop using Usemaps?
Quite the opposite. AI is redefining the relationship between humans and computers, as well as the way IT systems operate, while Usemaps is a tool that delivers the greatest value in human-to-human collaboration. It helps users focus on what truly matters – making decisions and developing concepts – instead of spending time on tedious, repetitive tasks such as searching for data, transforming it, or verifying its accuracy. As long as people work in teams, they will need tools that support and organize information and present it in a clear and meaningful way.
🔵 Will AI-powered tools be available in Usemaps?
They already are. Automatic vectorization of WMS layers with classification contours has been running continuously since early 2024 (although if initiatives such as those led by QGIS Polska are successful, we may eventually be able to abandon “re-vectorization” altogether, as data sources will simply become directly accessible). Data collected in Usemaps, when properly structured and described, can serve as a foundation for AI-driven insights (more on this in a future article).
🔵 Does AI support your coding process?
This is where the biggest hype is. LinkedIn is full of “examples” of working “applications” built in three evenings by someone who had never coded before. A quick script or a simple MVP – maybe. But Usemaps is a system that has been developed for nearly six years by a team of over a dozen people, with hundreds of thousands of lines of code. It is used by organizations managing critical infrastructure. There is no room here for writing code “on the fly” or relying on hallucinations.
Of course, we are closely observing this space. Every developer uses some kind of AI tool (we do not have a formal company-wide policy in this area), but the overall result is that… everything actually takes longer.
Why?
Because AI may generate code, the endpoint may pass initial tests, and then it goes to code review. And there, a senior developer with over 10 years of experience quickly sees that something could have been written in 50 lines of code instead of 250, often using overly complex or unconventional approaches. The code goes back for optimization. As a result, the amount of AI-generated code is actually decreasing, as it often turns out to be a false economy.
When building a system that will be developed and maintained for years, you cannot afford “spaghetti code” – it will come back to bite you. Sooner or later. Any future modification or refactoring will be significantly more expensive.
Perhaps this will change in the future, and we will be able to simply prompt: “Add a module to Usemaps that does X and Y.” But for now – not yet. Another factor is that there is still very little high-quality, publicly available code using modern GIS solutions. These AI systems simply have not had enough data to learn from. You will not find code on Stack Overflow that can process the entire BDOT10k dataset in a matter of seconds.