AI-first working on your own ground · For NGOs, political parties and government suppliers
You have to change anyway.
Make it a leap.
Moving away from American big tech feels like rebuilding without anything getting better. But if you have to change anyway, you can move forward in one go: reshape legacy processes, bring every existing connection into one intelligent AI layer and work in an environment with more overview and more control than today. On a sovereign stack.
From threat
to opportunity.
The same change, seen differently. Not from A to a European A, but from A to B.
- From switching
To improving. Not the same work in a different place, but working better than today.
- From separate tools
To one AI layer. Mail, files, CRM, member records and grant systems become reachable for AI, through one gate with fixed rules.
- From legacy processes
To processes that think along. What is now manual work across five systems, the AI prepares. People decide.
- From shadow AI
To AI with agreements. Staff already use ChatGPT or Copilot. Give them that convenience, with rules and a log.
- From trust
To proof. Every action of an AI agent is recorded: what, on whose behalf and with which data.
How an AI
workspace works.
See it as an operating system for your organisation: an AI-first Business OS. Think of an office suite. But instead of fixed file types, every file is its own app, built by AI for exactly your question.
An assistant that knows your organisation
One place where staff ask questions and hand off work in plain language. The assistant knows your terms, procedures, style and policies, and gets smarter as teams add their knowledge.
Your own apps, safely sandboxed
Ask for a dashboard, a member overview or a grant planning and the AI builds it. Every app runs in its own sandbox and can only reach what you give it. Share a good app with colleagues or turn it into a template.
A gatekeeper for every connection
Existing systems connect through gatekeepers. They only give access to what is intended, record every action and put anything that changes something before a human first. The AI keeps working; you approve or reject when it suits you.
You choose the model
A European model such as Mistral, an open model on your own servers, or another model for a task without sensitive data. Per task, not for everything. And it is in the log.
On European ground
The building blocks are open source and run on European servers or in your own data centre. No supplier who can pull the plug.
Those who must,
now can.
Three kinds of organisations face the same change. They mostly see the must. We show them the can.
NGOs
Must: reduce dependency on American services, because donors, partners and current events ask for it. Can: more impact with the same team, with grant applications, reports and campaigns the AI prepares.
Political parties
Must: member data is special category personal data, because it reveals a political opinion. It does not belong within reach of a foreign government. Can: branches, parliamentary groups and volunteers working with one assistant that knows the programme and positions.
Government suppliers
Must: governments want to know where data is and who can access it, and NIS2, the AI Act and Dutch government security baselines set requirements. Can: whoever has this demonstrably in order stands a better chance in tenders and can show it at any moment.
Overview.
Control.
Proof.
Strict rules are not a brake here. They are part of the design.
An AI workspace with gatekeepers records which agent did what, on whose behalf and with which data. That is what the GDPR, the AI Act and NIS2 ask of you: knowing what happens, and being able to show it.
We do not promise that software makes you compliant. People, agreements and evidence do. But we set it up so that the evidence builds itself, and sensitive data only goes to models you chose for it.
Curious where you stand now? Start with the frontrunner session: in one afternoon, the map of your digital dependencies, AI included.
Start small.
Improve
big.
Getting it right in one go does not mean everything at once. We start with the process that hurts most today.
- Step 1
Frontrunner session
€2,450 excl. VATOne afternoon with board, directors and key people: where do you stand, AI included, and what are the first three steps?
See the session - Step 2
One process, AI-first
After the session, tailoredOne legacy process, such as grant applications, member questions or reporting, moved to the AI layer. Small enough to learn fast, big enough to see the difference.
- Step 3
Connection by connection
After the session, tailoredExisting systems one by one behind a gatekeeper, with European partners. Set up so your own people can carry it afterwards.
Questions
we often hear.
What do you mean by an AI-first Business OS?
An operating system for your organisation instead of for your computer. Just as an ordinary operating system connects and separates programs, devices and users, a Business OS connects staff, AI agents, apps and existing systems, with fixed rules about who can access what. It is a way of setting things up, not one supplier's product. We assemble it from European, open building blocks that fit your organisation.
Do we have to move everything at once?
No. We start with one process that hurts and build from there. Existing systems keep running at first.
Does this work with our current systems?
Yes, that is the point. The AI layer connects to what is already there, through gatekeepers. So you replace a system only when that is better, not because you have to.
Can we still use American AI models?
You can, per task. For sensitive data you choose a European or your own model; for the rest whatever works best. You decide, and it is in the log.
Our people already use ChatGPT. What then?
Then you already know they want to work with AI. Your own workspace gives them that convenience, with agreements and without sensitive documents ending up with unknown parties.
First see the opportunity.
Then take the step.
Half an hour with Rik, no strings attached. We look at which process can get better first.
