Production systems, not pilots.
I work with mid-sized German companies and with public-sector and municipal organisations that want to adopt AI. The work is framed by outcomes, and each offer below links to a case study that shows it running in production.
Automate document-heavy and manual workflows
Invoices, reports and incoming post are opened, read and typed in by hand today. After this they are captured, validated and filed automatically, with deadlines tracked. Your people only see the cases where something does not add up.
Production AI over your own documents
Your own files become the source: contracts, reports, manuals, correspondence. A language model answers from them, and every answer is checked against the data your systems already hold. Built for daily use and for the people who work with it.
Adopt AI without giving up control
The worry about AI is rarely that it can do too little. It is that it gets something wrong and nobody notices. So I build systems where the AI reads and suggests, while the actual decision sits in code, where it can be read, tested and explained. Every result is checked before it moves on, the system would rather stop than guess, and anything that cannot be undone goes out only once a person has approved it. That way of working comes from fifteen years in an industry where a mistake is not a bug report but a recall.
Automate the workflows that still run by hand
Ordering, scheduling, updates across several systems: wherever someone carries data from one program into the next, errors and waiting time appear. The automation builds on the tools you already run instead of adding another silo.
Smart mobility for municipalities
A focus area I am actively developing, drawing on fifteen years in automotive systems engineering: AI and systems thinking applied to municipal mobility.
Stated as domain expertise and an initiative, not a delivered client engagement. There is no client case study for it yet.
01
Assessment
A short, clearly bounded engagement. We map the workflow, the data and the people involved, and record where AI genuinely helps and where it changes nothing.
02
Prototype
A working prototype against your real data, so the decision to go further rests on evidence rather than a slide deck.
03
Rollout and handover
Hardening, integration, measurement, documentation. At the end the system runs reliably and your own team can own it.
The way in is a short assessment.
The simplest way in is to scope one real workflow and find out where AI earns its place. Engagements run as a day rate or a fixed, project-based scope, whichever fits the work. If that matches your situation, write to me.