01
Check the outcome, not the setting
A backup schedule that's switched on doesn't prove anything. I check whether data actually arrives, and whether it can be restored.
Process automation & AI advice · Netherlands
AI in your business, step by step
No big promises, just results. Together we look at how work really flows today, find the places where automation and AI save time or prevent errors, and introduce them carefully — one process at a time.
Independent advice
You stay in control of your data
Built and tested myself
Three questions I always ask first
- 1What problem are we actually solving?
- 2What happens when the system gets it wrong?
- 3Could it be simpler — without AI?
AI: what it is and what it isn't
A lot is being promised. This is what I see in practice.
What AI is good at
- Summarising and drafting texts, emails and documents
- Finding information in your own documents
- Repetitive lookup and data-entry work
- Producing a first draft that a person finishes
What to watch out for
- It sounds convincing, even when it's wrong — review remains necessary
- Sensitive data doesn't belong in a cloud chatbot by default
- On a messy foundation, AI mostly makes the mess faster
- Not every problem is an AI problem — sometimes a simple script will do
“AI is a tool, not a strategy. The question isn't whether you use it, but where it demonstrably helps.”
A four-step approach
Starting from your business processes, not from the technology. Start small and only scale what works.
- 1
Map
How does the work really flow? Where is the manual work, the waiting, the re-typing? We look at practice, not at the org chart.
- 2
Choose
For each opportunity: what does it deliver, what does it cost, and what happens if it goes wrong? Sometimes the answer is a different agreement — not AI.
- 3
Start small
One process, with a measurable goal. The system does the groundwork; a person stays in charge.
- 4
Scale what works
What demonstrably works, we expand. What doesn't, we drop. Everything documented, so you can carry on yourself.
Services
From a first exploration to a working solution — and the foundation it needs.
Process analysis & AI exploration
Walking through your business processes together to find where automation really pays off. With honest advice, even when that advice is 'not yet'.
Step-by-step automation
Automating processes one at a time, with clear boundaries: what the system may do on its own, and where a person decides.
Moving to open alternatives
Want to move away from Windows, Microsoft 365 or the Apple cloud? I help with the switch to Linux, Nextcloud and other open, self-hosted software. Step by step, without losing data.
Privacy & local AI
Working with sensitive data? AI can run on your own hardware, without your data leaving the building. Set up with GDPR in mind.
Training & guidance
Teaching your team to work with AI assistants: what they're good for, how to check the output, and what not to put into them.
Reliable IT foundation
Backups that are proven to restore, a clear network and documented management, on Linux and open software. Without that foundation, AI makes little sense.
A deliberate choice: I no longer set up or manage Windows or Mac environments. If you want to move away from them, I'm happy to help.
How I work
Principles I apply in my own work every day.
Measure, don't assume
Advice is only worth something once it has been verified. That includes a vendor's recommendation — or an AI's.
People decide
AI may make suggestions. Anything irreversible only happens after a person approves it.
Your data, your control
I work with open software and self-hosted solutions. Only what is necessary and responsible goes to an external service.
Always a way back
Every change has a fallback: a tested backup, a snapshot, a plan B.
Documentation that lasts
You get explanations in plain language, so you don't depend on me.
Honest about limits
If something can't be done or isn't wise, you'll hear it. Better no promise than one that disappoints.
Lessons from practice
What assignments and my own environment have taught me recently.
02
One check, several reports
A technical report for internal use and a readable one for the client can come from the same check. Only what is genuinely human work stays manual.
03
Automate small and deliberately
Better one error-prone step automated well than a whole process automated halfway. Where real systems are affected, a person checks.
04
AI reads first, then asks
On client systems, AI may only look at first. Every change happens only after approval, and that approval applies per step.
05
Local is often good enough
For many tasks, data doesn't need to go to the cloud. A language model on modest hardware of your own is faster than you'd expect, and the data stays in-house.
06
Write down what didn't work too
A well-argued 'not chosen' saves a lot of time next time. Failed attempts belong in the documentation.
I'm happy to share concrete examples in a personal conversation. Schedule a call →
About Jean-Paul
I'm Jean-Paul Luijten and I've worked in IT my whole career. At Microsoft I was a Partner Technology Advisor (Cloud) for Windows Azure; as an independent I later worked as a cloud trainer, among other things, and gave talks on blockchain and AI. Explaining new technology and making it usable has always been the common thread. Precisely because I know Microsoft and Azure from the inside, I now help organisations go the other way: towards open software, where they stay in control of their own data.
These days I work with AI assistants every day. I run them on my own servers, so I also see what goes wrong. I also work for an IT service provider on self-hosted, privacy-friendly environments for its clients, so their data stays theirs. I bring that practice to your organisation: first understand how the work flows, then see what technology can add.
Business and IT
From account management to the server itself
Hands-on
I use what I recommend
Personal
Direct contact and short lines
Get in touch
Curious whether AI can do something for your organisation — or not? A first conversation is free of obligation.
Contact details
iSPART
Process automation & AI advice
Address
Groote Bolwerk 54
7202 AP Zutphen
Phone
+31 652 311 314KVK: 60887125
BTW: NL172782934B02
Start with a conversation
We look at your situation and your processes together. You get an honest picture of what automation and AI can do for you — and what they can't.
- No-obligation introduction
- Independent advice, no sales pitch
- Clear agreements up front