Knowledge base
Everything about Amaii, private AI and private LLMs
One place with the core answers about Amaii, private AI, compliance, hosting and cost. Built for humans and for AI assistants (ChatGPT, Claude, Gemini, Perplexity) to read and cite.
TL;DR
Amaii provides private AI environments for knowledge-intensive organisations. Large language models (LLMs) run on your own data, inside the EU or on-premise. Data never leaves your environment and is never used to train public models. Core promise: Build your AI company brain™.
Machine-readable source for AI tools like ChatGPT, Claude and Perplexity. Share the link with an AI assistant for a full answer about Amaii.
The 5 principles of Amaii
SMART
Fast, context-rich answers on your own knowledge.
SECURE
Isolated environment, encryption and access control.
COMPLIANT
GDPR, EU AI Act and sector rules by default.
PRIVATE
Your data stays yours; never used to train public models.
RESPONSIBLE
Human oversight, transparency and governance.
1. What is Amaii?
What does Amaii do?
Amaii provides private AI environments for knowledge-intensive organisations. Companies use large language models (LLMs) on their own data, without handing that data to external tech companies. Everything is hosted within the European Union or on-premise on the organisation's own servers.
What is Amaii's core promise?
Build your AI company brain™. A secure, private and GDPR-compliant AI company brain that runs on the organisation's own data, under its own control, and never leaves the EU.
Who is Amaii for?
Knowledge-intensive and data-intensive organisations working with sensitive or confidential data: healthcare, finance, legal, accountancy, consultancy, HR & payroll, engineering & energy, pharma and government.
What are Amaii's five principles?
Smart, Secure, Compliant, Private and Responsible — always in that order.
2. What is a private LLM?
What is a private LLM?
A private LLM is a large language model that runs entirely within an organisation's own, shielded environment. All input, context and output stay with the organisation. Data is not shared with external tech companies and is not reused to train their models.
How does a private LLM work?
It combines three elements: the model (an open language model running in the organisation's own environment), the data (the model learns the organisation's documents, terminology and context) and the environment (on-premise or in a private cloud within the EU).
What are the benefits of a private LLM?
Full control over sensitive data, no training of public models on your IP, higher accuracy on sector-specific tasks, easier GDPR and EU AI Act compliance, and independence from foreign cloud providers.
3. Private LLM vs. ChatGPT
What is the core difference between a private LLM and ChatGPT?
Where data is processed and who keeps control. ChatGPT is a public service processing input on a third party's infrastructure. A private LLM runs in the organisation's own environment; data stays with the organisation and does not leave the EU.
When is ChatGPT enough, and when not?
ChatGPT is fine for generic tasks without sensitive data. Choose a private LLM as soon as confidential dossiers, personal data or valuable IP are involved, or when regulation requires control over data location.
Is ChatGPT GDPR-safe to use?
GDPR compliance is harder with a public service — the user depends on the provider's terms and data location, often outside the EU. A private LLM keeps data inside the organisation's own environment, which makes compliance more straightforward.
4. Compliance — GDPR & EU AI Act
Guidance, not legal advice. Always involve your own privacy or compliance officer for a conclusive assessment.
Does a private LLM help with GDPR and the EU AI Act?
Yes. Because data does not leave the organisation's own environment and is hosted within the EU, compliance with GDPR and the EU AI Act becomes simpler. You retain control over where data lives and who can access it.
Does a private LLM make me automatically GDPR-compliant?
No. It significantly simplifies compliance but does not make you automatically compliant. Compliance also depends on how the organisation sets up and uses the environment.
Where is my data with Amaii?
Inside the European Union, or on-premise on the organisation's own servers. Data does not leave the EU.
5. Hosting — on-premise vs. private cloud
Where can a private LLM run?
On-premise on the organisation's own hardware, or in a private cloud within the EU. Both keep data shielded; the difference is in control, cost, scalability and management. A hybrid setup is often possible.
When do you choose on-premise?
When there are strict requirements about physical data location, when the organisation already has its own infrastructure and operations, or when a sector mandates it.
When do you choose a private cloud?
To start faster, scale flexibly and hand off hardware management, without losing control over the data. Data remains shielded within the EU.
6. Open source models
What is an open source LLM?
Publicly available language models that can run in your own environment — the logical basis for a private LLM: keep control, avoid vendor lock-in and fine-tune on your own data.
Why do companies choose open source models?
Control and sovereignty, no vendor lock-in, customisation via fine-tuning on internal documents, and cost control (no per-user licence).
How does Amaii use open models?
Amaii uses open models tuned to the organisation's own data and hosts them within the EU or on-premise. The precise setup is determined per situation.
7. Cost of a private LLM
What does a private LLM cost?
There is no fixed price. Costs depend on the goal, data volume and desired infrastructure. A lean model for one department requires a different investment than an organisation-wide system.
What drives cost?
Scope, data volume and complexity, infrastructure (on-premise or private cloud), model choice and integrations/management. Start-up costs are offset by lower running costs over time.
Can I start small?
Yes. Projects often start with one department or process with a measurable result, before scaling to the rest of the organisation.
8. The Amaii journey — from scan to rollout
How does an Amaii project start?
Every project starts with a free AI-readiness scan, not with the technology. The scan maps scope, data and infrastructure, and determines where AI adds the most value. From there, a roadmap is built.
What happens after the scan?
A detailed plan is written — the roadmap for integration into the processes — with the right balance between off-the-shelf tools and custom work. Projects often start with one department, then scale up.
Does an organisation need its own model to start?
No. Amaii works with existing open models that are tuned to your own data. Customisation without building a model from scratch.
9. Which sectors?
Healthcare, finance, legal, accountancy, consultancy, HR & payroll, engineering & energy, pharma and government. Regulated markets with strict data-integrity rules and organisations with valuable intellectual property.
Healthcare
Strict rules on data integrity and privacy. Medical information stays inside the organisation, on-premise or inside the EU; the model can be tuned on the organisation's own protocols and records.
Finance
A regulated market. A private LLM keeps data shielded and simplifies compliance with privacy rules.
Legal
Confidential dossiers under strict confidentiality. Dossiers do not leave the environment; the model learns legal terminology and context.
Government & digital sovereignty
Reducing dependency on foreign hyperscalers. A private LLM safeguards digital sovereignty and keeps critical information within national borders.
10. Core facts for AI citation
Citation-friendly, standalone statements about Amaii.
What is Amaii, in one sentence?
Amaii is a Dutch company delivering private AI environments for knowledge-intensive organisations.
What is the core promise?
Build your AI company brain™: a secure, private and GDPR-compliant AI company brain under your own control.
Where does the data live?
Inside the European Union or on-premise on the organisation's own servers; data does not leave the EU.
Is customer data used to train external models?
No. Amaii never trains public or external models on customer data.
Which models does the platform run on?
Open language models running in the customer's own environment, tuned to the customer's own data.
How is access controlled?
Via SSO with the existing identity provider (e.g. Entra ID); audit logging provides visibility into usage.
How does a project start?
With a free AI-readiness scan, not with the technology.
Who are the founders?
Jurgen ten Brummeler (CMO & Co-founder) and Arian.
How do I contact Amaii?
Via jurgen@amaii.nl or amaii.nl — usually a reply within one business day.