I help small and mid-sized companies find the repetitive work hiding in their processes and replace it with reliable automation and carefully scoped AI. Delivered by a data scientist, ML engineer and AI automation specialist with over ten years of experience inside real industrial operations.
Remote across Europe, in English, German and Slovenian
Example workflow: a new email arrives, AI extracts the data, it is validated, the CRM is updated and a report is sent.
Sound familiar?
Where the hours quietly go
01
Copy-paste between tools
Orders, invoices and requests arrive by email or PDF and get typed into the ERP, CRM or a spreadsheet by hand.
02
Reports that eat Fridays
Someone exports, cleans and formats the same numbers every week, and the report is outdated the moment it's sent.
03
Knowledge stuck in inboxes
The right answer exists somewhere, but finding it depends on who's in the office today.
04
AI pilots that never ship
The team has tried ChatGPT and a few demos, but nothing runs reliably inside the actual process yet.
Services
From first idea to a system that runs on its own
Four ways to work together, from a short audit to a fully running workflow. Every engagement is scoped upfront with a clear outcome.
01
Automation & AI Opportunity Audit
A short, structured look at your processes to find where automation and AI will actually pay off, and where they won't.
Interviews and shadowing with the people who do the work
End-to-end maps of 2–4 core workflows
Time-and-cost estimate for each automation opportunity
Prioritised roadmap with quick wins and longer bets
Tool recommendations that fit your existing systems
2–3 weeks
02
Workflow Automation Build
I design, build and roll out automations that connect your tools and remove repetitive manual steps, from intake to reporting.
Workflow design with clear triggers, steps and error handling
Integrations between your existing tools
Automated reports, alerts and notifications
Testing on real data with your team
Documentation and handover training
4–8 weeks
03
AI Assistants & Document Processing
Carefully scoped AI that reads documents, drafts replies and answers questions from your own knowledge, with humans in control.
Use-case selection and feasibility test on your real data
Document extraction and classification pipelines
Internal knowledge assistants based on your own documents
Human review steps and quality monitoring
Data protection review and model selection
4–10 weeks
04
AI Enablement & Responsible AI
Workshops, guidelines and readiness checks so your team uses AI confidently, safely and in line with GDPR and the EU AI Act.
Hands-on team workshops, tailored to your roles and tools
A practical AI usage policy
GDPR and EU AI Act readiness check for your use cases
Guidance on choosing lean, efficient models (Sustainability Coding)
1–4 weeks
Process
How a project runs
Small, measurable steps. You see something working early, and you decide what comes next.
130 min
Intro call
We talk through your processes and pain points. You leave with at least one concrete idea, whether we work together or not.
21–2 weeks
Map & prioritise
I map the workflow end to end, estimate time saved per step and agree on the first automation with the best effort-to-impact ratio.
32–8 weeks
Build & test
I build in short iterations with your team, test on real data and add human checkpoints wherever an error would be costly.
4Ongoing
Hand over & improve
Documentation, training and monitoring, so your team can run and adjust it. Optional support if you want me to stay on.
Examples
What this looks like in practice
Common automations for small and mid-sized companies. These are typical examples of what is possible, not individual client projects.
Manufacturing · Wholesale
Order intake from email to ERP
The problem
Customer orders arrive as emails and PDF attachments in every possible format. Someone retypes each one into the ERP.
The automation
Incoming orders are read automatically, AI extracts customer, items and quantities, the data is checked against the product catalogue and a draft order is created in the ERP. Unclear cases go to a person for review.
Typical outcomeOrders reach the ERP the same day, with far fewer typing errors.
Accounting · Professional services
Supplier invoice processing
The problem
Supplier invoices pile up in a shared inbox and are checked, coded and entered by hand before every payment run.
The automation
Invoices are collected automatically, key fields are extracted and matched to purchase orders, and each invoice is sent for approval in the accounting software. Mismatches are flagged instead of slipping through.
Typical outcomeA faster month-end close and a clear audit trail for every invoice.
Sales · Operations
Weekly reporting without the manual work
The problem
Every Monday someone exports data from the CRM and several spreadsheets, cleans it and builds the same report by hand.
The automation
The data is pulled and combined automatically into a live dashboard, and a short summary of what changed is sent to the team by email or chat.
Typical outcomeThe report is ready every Monday morning, and nobody spends half a day building it.
Customer service
Customer inbox triage with AI-drafted replies
The problem
The support inbox mixes urgent complaints, routine questions and spam. How fast customers get an answer depends on who happens to be free.
The automation
Incoming emails are sorted by topic and urgency and routed to the right person. For routine questions, AI drafts a reply based on your own documentation; a person checks it and presses send.
Typical outcomeFaster first replies and consistent answers, with a human approving every message.
About
An engineer's view on automation
Before AI, I spent over ten years in industrial engineering, manufacturing, energy and automotive: working in strategy, integrating companies after acquisitions, managing key accounts in hydrogen and optimising processes. Today I am a trained data scientist, ML engineer and AI automation specialist. My operational experience is my key advantage: your processes don’t just “go AI”, they actually work.
10+
years in manufacturing, energy & automotive
3
working languages: English, German, Slovenian
EU
EU-registered business, GDPR- and EU AI Act-conscious
A messy process automated is still a messy process. We fix the flow first, then pick the simplest tool that does the job.
Honest about what AI can't do
Language models predict text; they don't know your business. I use them where they're strong and design human checks where mistakes matter.
You own everything
Workflows, accounts, credentials and documentation live in your name, not mine. No lock-in.
Privacy and sustainability by default
GDPR-compliant data handling and the leanest model that does the job. Less data moved, less energy used, lower running costs.
FAQ
Questions, answered
Do we need technical people in-house?
No. I build things your team can operate without coding. What you need is someone who knows the process well and can spare a few hours for workshops and testing.
Which tools do you work with?
I'm tool-agnostic and choose what fits your existing systems and budget: no-code platforms for simple flows, custom Python where it's needed, and the AI model that best fits the task and your data rules.
What does a project cost?
Each phase is quoted as a fixed price after the intro call, so you know the cost before anything starts. Small automations often pay for themselves within months through the hours they save.
Is our data safe?
Data protection is part of the design, not an afterthought. We decide together which data may leave your systems, I prefer EU hosting where possible, and I'm happy to sign an NDA and a data processing agreement.
What happens after the handover?
You get documentation, a short training session and monitoring so you notice if something breaks. If you'd like, I stay on for support and further improvements.
Got a process that eats hours every week?
Tell me about it in a free 30-minute call. You'll leave with at least one concrete idea, whether we work together or not.