Internal tools
Tools written to be read after we leave
Internal tools get a bad reputation because they are usually built by whoever had time that week, with no tests and no documentation. Then the person who built it leaves, and the next hire inherits a black box.
We version and review them the same way we version and review commercial software. A data-entry tool that crashes on malformed input costs more in staff hours than a proper input-validation pass costs to write — and scope is usually small enough to fit inside a few weeks.
Windows · internal network · weeks, not months
Customer-facing utilities
Distribution is part of the product
The user has no obligation to be patient. If the installer fails or the app looks untrustworthy, they uninstall and never come back. Features will not save a product that trips a security warning.
Code signing, a clean installer, an update mechanism and a clear uninstall path are part of the deliverable, not afterthoughts. We handle Windows desktop distribution directly and we do not sell app-store publishing as a standalone service.
Signed build · update channel · uninstall path
Simulation front-ends
The interface behind the model
A meaningful share of our desktop work sits behind a simulation. The engine handles the physics or the process model; the desktop application handles the interface, the session management, the scoring and the reporting.
This is where the game background earns its keep. Trainee-facing software has to onboard people who are not technical users, and the techniques for that are the same ones a game uses to teach a mechanic without a tutorial wall. Belgian studios in this space, LuGus Studios in Genk among them, work the same overlap from the game side.
Session control · scoring · reporting
Crossover projects
One pipeline, not two teams
A VR training module is a game-engine build with a business software requirement attached. A simulation with a desktop companion app needs both at once. Those are the projects we get called in for most often.
For those we run a single pipeline rather than splitting the team in half. The build system, versioning and review process stay the same, and the deliverable is decided per project rather than per discipline, so crossover work never gets treated as a special case.
VR · desktop companion · shared build system
Scoping & handover
Organised around the user, not the feature list
Scoping starts with the person who will run the thing. Who operates it, on what machine, in what context — and what happens on the day it breaks? From there we write a brief covering the platform target, the distribution method, the update path and the acceptance criteria.
Acceptance criteria are the part most clients skip, and the part that stops an argument at the end of the project. If the tool must talk to existing systems — a database, a sensor, an internal API — that gets identified during scoping rather than discovered during testing. Handover means source code, build instructions, and a walkthrough with whoever will maintain it.
Source, build instructions, walkthrough session