AI · AUTOMATION · TRAINING
I help people understand and make practical use of AI and automation. I combine process thinking with technologies that can reduce routine work — without hype or unnecessary shortcuts.
01 / STORY
I had already started simplifying things before that. AI is another tool to me, not a passing trend.
Automotive
I spent several years working with FMEA and risk analysis. It taught me to look at processes through possible failure modes, their effects and prevention, and to facilitate expert work in a structured way.
Self-directed learning · move into IT
I moved into IT partly through self-directed learning. In SAP Personas, I then built automations that removed dozens of manual steps for users.
Generative AI · knowledge sharing
I began testing the new possibilities of AI intensively and showing them to colleagues. This gradually led to a training role and the opportunity to give public talks about AI.
AI · API · custom components
I connect tools that do not normally communicate and turn them into solutions that reduce manual work. When a ready-made component is missing, it can be programmed and integrated into the whole.
02 / WHAT I DO
Understand the problem first, then choose the tool. Technology is a means, not the default answer.
From the first prompt to automation and agents. I explain the principle, demonstrate it with an example and let participants use the technology themselves.
BEFORE IMPLEMENTATION BEGINS
Map the process, data, permissions and risks; verify whether AI or automation makes sense; and recommend a realistic next step, including a smaller proof of concept.
BUILD · TEST · UNDERSTAND
I do not build technology for its own sake. I connect specific tools where they can free people from routine work.
03 / HOW I THINK
A working path is only the beginning. I want to know what happens when reality stops being ideal.
What if a step does not happen at all, happens only partly or happens at the wrong time?
What if the incoming data is incorrect, incomplete or unexpected?
What if the user or a connected service behaves differently from what we assumed?
How can we detect the problem early, limit its impact and keep the solution maintainable?
I apply my FMEA experience to automation design. I consider not only the ideal path but also possible failure modes, their effects, detection and prevention. I try to break a solution deliberately before an end user discovers the fault.
04 / PROJECTS
Specific situations where technology removes routine work, connects data or helps verify what is genuinely useful.
The same information was entered in Jira and then again in a Word report. This created duplicate work and often incomplete tickets.
The automation takes the data from Jira and prepares the output document. When standard tools were not enough, I added a custom Python component.
The automation collects news from a precisely defined field, evaluates relevance, reduces duplicates and prepares a summary of the previous 24 hours.
The result is up-to-date information without having to browse a large number of sources and posts manually.
I test different forms of agentic workflows — in workflow tools, cloud services and on my own hardware. I explore when a task can be given in natural language, for example from a phone or chat, and an agent can carry out the subsequent steps.
The goal is not to chase every new development, but to understand where an agent genuinely simplifies work or everyday life and where it merely adds another layer of complexity.
Example: an agent in a self-hosted environment can extract the required values from a screenshot of a running app and record them in a spreadsheet without manual retyping.
05 / CONTACT
Are you considering training, automation or a specific process? Get in touch. We will first clarify what is actually worth addressing.