What has changed
A broadly positioned software and web development practice has, over time, grown into something clearer: an engineering team that builds AI systems for mechanical and plant engineering. Operating manuals, service reports, drawing archives, twenty years of costing data - knowledge these companies have in abundance, just rarely where the decision is being made. What we build for our clients every day is also what this site is meant to represent.
Until now there was a considerable gap between what we do and how we presented ourselves. We closed it by rebuilding the site from scratch - with clearly delineated services, an open methodology, and three anonymised case studies that describe our work as it actually was.
What you will find here
We are not going to write about the ten most important AI trends of 2026 here. Nor are we going to comment on every model release.
What you will find instead are notes from our own work: architectural decisions that held up - and ones we had to correct. Economic considerations before and after go-live. Open source contributions that come out of specific projects. Occasionally an argument we would like to discuss.
The first of them are three MCP servers we have released as open source: for MOCO, for sevdesk and for Outlook. Three tools we use internally ourselves - and that we missed back when they did not exist.
What we would like
If you read something on these pages that is wrong, unclearly put, or does not match your experience - write to us. We would rather improve than justify.
And if you are currently facing an AI decision and are not sure whether what you are being offered fits the reality of your operation - give us a call. 30 minutes. No pitch.
