Our data is a mess.
Sharpen the data foundation and use case first, then the model. Never the other way around.
AI & Data Consultant for mid-sized business (DACH)
Where business meets
AI.
Business background, tech-fluent. 10 years of Business Intelligence, focused on GenAI since 2023. Pragmatic solutions for mid-market and startups across the DACH region.
Book a meeting Currently fully booked — inquiries for Q3/Q4 2026 welcome.
- GDPR-compliantfrom day one
- EU AI Actaccounted for
- EU hostingby default
Selected clients & engagements
Pharma & Nutraceuticals CDMO 
The market is shifting
AI has arrived in mid-sized business. Faster than expected.
The numbers are clear. Start now and you build a lead. Wait, and you pay to catch up later.
$4.4T
in potential annual value from generative AI. Worldwide, across all industries (range $2.6T–$4.4T).
Source: McKinsey, 2024Where projects fail
I spot the bottlenecks before they become a problem.
AI projects rarely fail on the technology. They fail on messy data, missing strategy and unclear value. I see these bottlenecks in almost every project. I solve them before they cost budget and time.
Only26%
of companies generate real business value from AI. The rest stall in the pilot.
Source: BCG, 2024The right AI solution for every challenge.
We don't know where to start.
An assessment prioritises the use cases with the biggest business leverage.
Our last AI project fizzled out.
A small, controlled PoC with real data. We only scale once it delivers.
We need a real solution, not slides.
From idea to working software. Built myself or scaled with trusted partners.
Fluent at the intersection of
business and technology.
Trained in business administration, picked up the technical skills on the job — which grew into 10 years of Business Intelligence and a perspective that bridges both worlds. Since 2023, my focus has been on GenAI.
I'm currently supporting a leading pharma and dietary supplements manufacturer with AI-driven process optimisation — and an automotive startup with the build-out of its BI landscape.
My compass is business impact. I prioritise honestly, even when that means scoping a supposed AI project down to its essentials. I speak the language of the CTO as much as the CFO, don't get lost in technology for technology's sake, and ship working solutions instead of strategy decks.
When I'm not working with clients, I build my own automations and experiment with new LLM releases.
Projects & Delivery
Three formats, depending on the maturity
of your AI initiative.
Whether it's an initial assessment, a validated use-case or ongoing delivery — I step in where the leverage is highest.
01
AI & Data Assessment
Status quo, opportunities, prioritised roadmap — before a single euro flows into implementation.
- Interviews with business and IT stakeholders
- Analysis of data and tool landscape
- Use-cases prioritised by business impact
- Implementation plan with concrete next steps
2–4 weeks
02
Proof of Concept
Move one prioritised use-case from paper to a working prototype.
- Technical concept and architecture
- Hands-on development (GenAI, RAG, automation)
- Evaluation with real data and key users
- Decision basis for rollout or exit
4–8 weeks
03
Embedded AI Lead
Operational delivery and strategic guidance as part of the team.
- Technical lead for ongoing AI initiatives
- Sparring with management and stakeholders
- Coaching internal teams to build capability
- Bridge between business, data, and engineering
3–6 months · 1–3 days / week
Delivery & software development. What can be built in-house, I build myself. Larger scope I scale with a network of trusted partners.
Daily rates from €1,000 — final scope and model (fixed price or T&M) defined per engagement.
Workshops & Enablement
Bring AI into the team,
not just the project.
Sometimes the biggest leverage isn't a project — it's a team that uses AI with confidence. Two formats, both hands-on with your real tasks.
Flexible scope — from a compact single day to a full week, depending on your team and how deep you want to go.
01
AI foundations: working with LLMs
From cautious experimenting to confident day-to-day work with language models — model-agnostic, not tied to one vendor.
- Get your bearings across the model landscape: which LLM fits which job (ChatGPT, Claude, Gemini and beyond)
- Practical prompting that holds up in daily work, built around your own tasks
- Concrete use cases for each team
- Data protection and guardrails: what you can share, and what you can't
Outcome
A team that uses LLMs with confidence.
02
Automating processes with AI
Automate recurring workflows with AI. The result is a running flow, not a slide deck.
- Spot where your processes can be automated
- How AI automation actually works, using tools like n8n or Make
- Plug in AI building blocks: LLM calls, documents, and data
- A first working automation, built live in the session
Outcome
A first automation that actually runs.
Tailored to your team. Also as a combination of both tracks or a series across several sessions — from first steps to building internal champions.
Workshops run on a fixed price, not a day rate — preparation, materials and tailoring to your team included. Price depends on format and scope.
How we work together
Structured from idea to delivery.
Every project is different. My approach is proven. Clear phases, measurable milestones, full transparency.
- 01
Get clarity
Analysis of processes, data and potential. We find the use cases with the biggest leverage.
- 02
Prove it
A controlled proof of concept with real data. We only scale once it delivers.
- 03
Build it
Implementation and software development in short cycles. What I can build myself, I build myself. Larger scope I scale with trusted partners.
- 04
Embed it
Integration into operations, team enablement, measurable results tracking.
Fixed-price optionFull transparencyNDA & IP protectionRegular updates
Data protection is not an add-on.
Every AI solution is built data-protection-compliant from the start.
GDPR and the EU AI Act are designed in, not bolted on. EU hosting is the default.
FAQ
Frequently asked questions.
Are we too small or too early for AI?
Rarely. The right entry point is a question of leverage, not size. Often an assessment or a compact workshop is enough to find a first worthwhile use case at low risk.
What if our data isn't ready yet?
That's the norm, not a dealbreaker. I sharpen the data and the use case first, then bring in the model. You don't need a perfect data foundation to start — just an honest look at it.
How do we keep it from stalling at a prototype?
Start small, deliver early, only scale what holds up. I build working software instead of slides and embed it into day-to-day operations. A PoC that proves nothing gets stopped deliberately, not kept on life support.
How quickly will we see results?
Often faster than you'd expect. An assessment usually delivers a prioritised roadmap in 2 to 4 weeks, a proof of concept a working prototype in 4 to 8 weeks — and noticeably sooner for a clearly scoped use case. A workshop works immediately: your team walks out the same day with new skills.
Do you also offer workshops and training?
Yes. Two formats: AI foundations with language models (model-agnostic) and process automation with AI, both hands-on. Fixed price, tailored to your team — see the Workshops section for more.
What does an AI project — or a workshop — cost?
Projects start at a day rate of €1,000, on a fixed price instead of time and material if you prefer. Workshops always run on a fixed price, preparation included. The final scope depends on the format.
Is the solution GDPR-compliant?
Data protection is designed in from the start, not bolted on. Development follows GDPR and the EU AI Act, EU hosting is the default, and data processing agreements and data flows are clarified right at the start of the project.
Working directly with you or through an agency?
Directly, with no layers in between. Around 10 years of Business Intelligence plus a GenAI focus since 2023, hands-on in delivery. Larger scope I scale with a network of trusted partners.
Let's talk.
Have a specific question, or want to gauge informally whether a collaboration might fit? A 30-minute call is the fastest way.
Book a meeting Currently fully booked — inquiries for Q3/Q4 2026 welcome.
Or directly