VANDAMPrivate AI
Sensitive steps on your own modeland Claude for the rest.
We map which steps of a workflow touch data that should stay with you, run an open-weight model on your hardware or an EU host for those steps, and route the rest to a frontier model such as 
Claude. Tests on your own cases show where the local model is good enough.
$16,000 three weeks · plus hardware or hosting at cost
Plan Sam’s return after six weeks of sick leave.
Six steps in the HR workflow at Atlas & Co. Steps that touch health or personnel data run on a model inside the building. The rest go out to a frontier model.
- 1Read the doctor’s noteHalf days from November 9
- 2Pull the absence recordFirst absence in three years
- 3Remove names and the diagnosisEmployee A, condition withheld
- 4Check the collective agreementPhased return up to 8 weeks
- 5Draft the return planFour-week plan, drafted
- 6Put names back and file itFiled, manager told
- Stayed inside
- 1.86 MB
- Left the building
- 7.7 KB
Four of six steps ran on the company’s own server. Only text without names or health data left the building.
| Step | Local | Frontier |
|---|---|---|
| 1 | 95 | 97 |
| 2 | 99 | 99 |
| 3 | 97 | 98 |
| 4 | 72 | 94 |
| 5 | 81 | 93 |
| 6 | 98 | 99 |
Privacy holds back AI in Europe.
Nearly half the companies that considered it say so.
Dell for private AI (AI in a box for firms that won’t use the cloud)
Late CheckoutOctober 2026Source Most companies we meet are fine with the cloud for most of it. Just not for the sick notes.
Most of a workflow can use the cloud.
A few steps should stay home.
A sick note, an employment contract, a client file in a dispute. Some data should not go to an outside model, because of the law, a works council, a client contract or plain judgment, and many AI projects stop right there. Yet in most workflows only a few steps touch that data. Reading the doctor’s note is one of them. Checking what the collective agreement allows is a different kind of step.
Open-weight models can now run those steps on a server you control. Aleph Alpha released Kolibri in October 2026, built in Europe, with weights under Apache 2.0 to run on your own hardware. A month earlier 
Perplexity let its Computer start a task in the cloud and move the steps with private files to a local model on a Mac. 
Y Combinator asks for startups that help companies own their intelligence, and for AI for the air-gapped economy.
We split the workflow by the data each step touches. Sensitive steps run on an open-weight model in your building or with an EU host. Names and identifiers are removed before anything goes out, and the remaining steps go to a frontier model such as Claude, which still does the hardest reasoning best. Every step is tested on your own cases with both models, so the split rests on measured scores. The model, the routing and the tests are yours, as described in own your intelligence and on our models page.
In June 2026 @ryiacy described where companies that hand their AI work to outsiders end up: paying for the training of the model that replaces them, unless they “find a way to build and own their own model ecosystem.” Running the sensitive steps yourself is a practical first part of that.
What IT and HR look at after the handover.
A routing view per workflow: every step, the data it touches, where it ran and how much left the building. Anyone can check that the doctor’s note stayed on the local server and that only text without names went out.
Next to it sit the scores per step for the local and the frontier model on your own test cases, with the bar your team set. When a new open-weight model ships, it runs against the same cases, and a step moves inside once it clears the bar.
What every Private AI setup contains.
Your workflows and hardware differ. The setup is built the same way each time.
Sensitivity map
Every step of the workflow and the data it touches.
Model choice
Open-weight models tested on your cases, the best fit chosen.
Local install
The model running on your hardware or with an EU host.
Routing rules
Which step runs where, written down and enforced.
Redaction
Names and identifiers removed before text goes out.
Eval comparison
Local and frontier scores per step on your own cases.
Routing log
Where each step ran and how much data left.
Runbook
Updates, outages and new workflows, step by step.
How the split gets made.
Map the data in each step
We walk through the workflows with the people who run them and with your lawyer or data protection officer. Each step is marked by the data it reads and writes: health, personnel, client or public. The map says which steps stay inside and why.
Install the local model
We choose an open-weight model that fits the job and your hardware, and run it on a server in your office or with an EU host. Routing sends each step to the right model, and names and identifiers are removed before text goes out.
Test both, then hand over
Every step runs on your own cases with the local model and with a frontier model such as
Claude. The scores show where local is good enough. You get the setup, the routing rules, the tests and a runbook for updates, outages and new workflows.
Set up once.
More steps move inside as models improve.
Open-weight models get better every few months. After the handover we can test each new release on your cases, update the local model when it clears the bar and move more steps inside as the scores allow.
Hosting
The local model kept running, patched and monitored, on your hardware or with the EU host.
Model updates
New open-weight releases tested on your cases before they replace the current one.
Routing review
The routing rules and the log of what left the building checked every month.
Review
A monthly session with IT, HR and legal on what moved inside and what comes next.
What you
receive.
Sensitivity map
Every step, the data it touches and where it may run.
Hybrid setup
A local model, routing rules and a log of what left the building.
Eval comparison
Scores per step for the local and the frontier model, plus a runbook.
Where Private AI leads.
Each next step has its own service.
Before you ask.
What hardware do we need?
For one or two workflows, usually one server with a recent GPU, in your office or rented from an EU host. We size it on your volumes in the first week. Hardware or hosting is billed at cost.
Why not run everything locally?
Frontier models such as Claude still reason better on the hard steps, and most steps in a workflow touch nothing sensitive. The tests show per step where a local model is good enough.
Does this make us GDPR compliant?
It keeps health and personnel data on infrastructure you control and logs what goes where. Your lawyer or data protection officer makes the legal call, with the sensitivity map and the log in hand.
Which open-weight model do you use?
The one that scores best on your cases and fits your hardware. We test a few, European models among them, and you can switch later without rebuilding the routing.
Who owns the setup?
You do. The model runs on your hardware or your account with the host, and the routing rules, tests and runbook are yours.