Production systems, not pilots.
I take on selective independent engagements with German Mittelstand companies and public-sector and municipal organisations adopting 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
Turn paper, inboxes, and manual data entry — invoices, reports, intake — into structured, reviewed data with deadlines tracked. The repetitive part runs unattended; people handle the exceptions.
Production AI over your own documents
Extraction and retrieval pipelines that answer from your files, not the open web — classified, validated, and reconciled into data your systems and staff can actually use. Built to run in production, not to demo once.
AI adoption with the rigor of a regulated industry
Deterministic decisions, validation at every boundary, fail-safe defaults, and a human in the loop for anything irreversible — the discipline of fourteen years in safety-critical engineering, applied to LLM systems. For risk-averse organisations, this is the difference between a pilot and something you can trust in production.
Process automation
Replace manual, repetitive, error-prone processes — coordination, ordering, multi-system updates — with reliable automation that integrates the tools you already run, rather than adding another silo.
Smart mobility for municipalities
A focus area I am actively developing, drawing on fourteen years in automotive systems engineering — applying AI and systems thinking to municipal mobility.
01
Assessment
A short, focused engagement to map the workflow, the data, and where AI genuinely helps — and, just as important, where it does not.
02
Prototype
A working prototype against your real data, so the decision to go further rests on evidence rather than a slide deck.
03
Production & handover
Hardening, integration, measurement, and documentation — so the system runs reliably and your own team can own it.
Start with a short assessment.
The cleanest way to begin 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 fits what you are facing, get in touch.