---
title: "AI Consulting on Open Models in Switzerland"
description: "AI consulting for Swiss companies on open-weight models, hosted in Switzerland or on-premise. Data sovereignty, FADP, GDPR and FINMA compliance, no vendor lock-in."
url: "https://ai.malagoli.me/en"
locale: en_US
alternates: [it-CH, fr-CH, de-CH, en]
---

# Artificial intelligence that stays home.

<!-- AI consulting · Open models · CH -->

We design AI solutions built on open models, hosted wherever you decide. Frontier-model capability, without handing your data to foreign clouds — and in full compliance with the FADP, GDPR and FINMA guidelines.

- [Book a call →](#contatti)
- [Why open models](why-open-models.md)

_Free initial discovery call · No obligation_

```yaml
data_residency: CH ✓
FADP:           compliant ✓
lock-in:        none
models:         open-weights
```

## Most AI solutions send your data overseas.

Every time a company uses a model via a US vendor's API, its data — contracts, customer data, financial information, medical records — leaves the infrastructure it controls and ends up on servers subject to foreign jurisdictions. For many Swiss companies this isn't a technical detail: it's a legal, reputational and compliance risk.

> The right question isn't "is the AI powerful?" It's: **where does my data end up, who can access it, and am I still compliant with Swiss law?**

## A different way to do AI in business.

### 01 · Your data stays where you decide.

We build on open models running on your own infrastructure or on a Swiss sovereign cloud. No data sent to third-party APIs. No training on your content. No gray areas.

### 02 · You own the solution, you don't rent it.

Open models are yours to run, portable and inspectable. No dependency on a single vendor, no pricing surprises, no features disappearing overnight. If the market shifts tomorrow, you keep running.

### 03 · Compliant by design.

Every solution is designed around the FADP, GDPR and — where needed — FINMA guidelines and sector-specific requirements. Data residency in Switzerland isn't an add-on: it's the starting point.

## From strategy to deployment, under your control.

## Real experience, not slides.

ai.malagoli.me was born from years of work on enterprise IT infrastructure, security and applied AI projects in regulated contexts — from finance to industrial. We don't sell the latest tech fad: we build systems that have to work, pass an audit, and hold up over time.

`Experience in FINMA-regulated environments` · `AI projects in production` · `Background in enterprise security and infrastructure` · `100% data in Switzerland`

## A four-step path, with no surprises.

| #  | Phase                 | What happens                                                  |
| -- | --------------------- | ------------------------------------------------------------- |
| 01 | Discovery             | We understand your context, regulatory constraints and where  |
|    |                       | AI brings real value.                                         |
| 02 | Proof of Concept      | A working prototype on a concrete case, so decisions are made |
|    |                       | on facts, not promises.                                       |
| 03 | Deployment            | Secure production rollout, on your own infrastructure or on a |
|    |                       | Swiss cloud.                                                  |
| 04 | Governance & Handover | Policies, documentation and training so the system remains    |
|    |                       | yours.                                                        |

## The right objections.

### Are open models less capable than closed ones?

On some highly complex tasks, the best closed models remain ahead. But for the vast majority of business use cases, modern open models are more than sufficient — and the advantage of keeping your data under control far outweighs a few points of performance. Where it genuinely makes sense, we adopt a hybrid approach.

### Where is my data hosted?

Wherever you decide: on your company's own infrastructure, on-premise, or on a Swiss sovereign cloud. Sensitive data is never sent to foreign vendors' APIs.

### Do we need expensive infrastructure?

It depends on the use case. Some solutions run on modest hardware; others require dedicated GPUs. During the Discovery phase we work out the cost-benefit trade-off together, with concrete numbers before any investment.

### Are we compliant with the FADP and GDPR using this solution?

We design every solution around the requirements of the FADP, GDPR and — where applicable — FINMA. Formal compliance should always be validated with your own legal counsel, but we start from an architecture built to meet them from the outset.

### How long does a project take?

A proof of concept can take just a few weeks; a full deployment depends on complexity. Our phased method exists precisely to give you visibility and control over timeline and cost, step by step.

### Do we stay dependent on you?

No — that's a principle for us. We hand over documentation, policies and training so your teams can run the system independently. Our goal is to make you self-sufficient.

## Let's talk about your situation.

Thirty minutes to figure out whether AI on open models is the right choice for your company — and what it would actually take. No sales pitch, just an honest technical conversation.

- [ ] First name
- [ ] Last name
- [ ] Company
- [ ] Email
- [ ] Short introduction — _A couple of lines about you, your company and what you need._
- [ ] I have read and accept the privacy policy for the processing of my data for this request. *

**Send request →** · [info@ai.malagoli.me](mailto:info@ai.malagoli.me)

> response within 1 business day · data processed in accordance with the FADP

why-open-models.md

# Why we build on open models.

Not for ideology. For a concrete reason: on sensitive data, who owns and controls the model matters more than a few points of performance on a benchmark.

## "Open" models, in brief.

An open (or open-weight) model is an AI model whose weights — the trained "brain" — are publicly available and can be run on infrastructure of your choosing. The alternative is closed models, accessible only through a vendor's API: powerful, but "black boxes" you have to send your data to in order to get an answer.

> With a closed model, the data goes to the model; with an open model, the model comes to the data.

## Four reasons this matters.

### 1 · Data sovereignty

With an open model running internally, sensitive data never leaves your perimeter. No sending it to foreign APIs, no transit through cloud infrastructure outside Switzerland, no risk of your content feeding the training of a third party's model.

### 2 · No lock-in

A model you own is portable and stable. You don't depend on one vendor's decisions: price increases, changed terms, deprecated features or sudden service interruptions don't expose you. The system you build today remains yours tomorrow.

### 3 · Transparency and control

Open models are inspectable. You can verify their behavior, adapt them to your domain, subject them to security review and document them for an audit. In a regulated context, being able to explain _how_ a system works matters just as much as the fact that it works.

### 4 · Economics at scale

Above a certain volume of usage, owning your own infrastructure becomes more predictable and more economical than a per-call cost that grows without limit. The costs are yours, controllable and plannable.

## What about closed models? Let's not pretend they have no advantages.

It would be dishonest to say open models always win. On some particularly complex tasks, the best closed models remain ahead today, and adopting them is immediate because it requires no infrastructure.

That's why our approach is pragmatic, not dogmatic:

The point isn't "open versus closed." It's putting every piece of data where it belongs.

blog/index.md

# Practical notes on AI, data and compliance.

Short, concrete articles on open models, RAG and FADP/GDPR/FINMA compliance — written for people who have to decide, not just read about AI.

llms.txt

# ai.malagoli.me

> AI consulting on open models for Swiss companies. We design AI solutions hosted in Switzerland (on-premise or sovereign cloud), built on open-weight models, compliant with the FADP, GDPR and FINMA guidelines, without sending sensitive data to foreign APIs and without lock-in to a single vendor.

Areas of activity: strategy and AI readiness, selection and deployment of open models, RAG systems over company documents, automation and AI agents, AI governance and compliance, team training. Based in Ticino, Italian-speaking Switzerland. Contact: info@ai.malagoli.me

## Main pages

© 2026 ai.malagoli.me · data hosted in Switzerland · CH ✓
[Privacy Policy](/en/privacy) · [Cookie Policy](/en/cookie-policy)