SECURITY AND COMPLIANCE
Security, data and EU regulation
INTENTIA+ keeps what your organization decides with AI. This page explains where that data lives, who can do what, and how you check that the record is intact.
- Where Servers in Europe
- Regulation Complies with the EU AI Act
- Your AI Your account, your contract, your terms
- Deciding Always a person, with a second factor
Where your data lives
- In Europe. INTENTIA+ runs on servers in Europe.
- Your text goes only to your AI provider. With your OpenAI, Anthropic or Google account and under your terms. So does whatever the assistant searches on the internet, which goes through that same provider's search.
- The guides, separately. When you ask AICA how INTENTIA+ works, the question travels with fragments of the user guides and nothing from your workspace.
- No cookies and no third-party tools. What AICA learns are usage facts, with no text and without saying who. The only text of yours that reaches the INTENTIA+ team is the question you choose to send it, after seeing it.
Who can do what
- Isolation per workspace. Each workspace is separate. Permissions are checked on every request, and if the database does not answer, nobody gets through.
- Roles. Who administers, who edits, who reviews and who only reads. Each person sees and does what their role allows.
- Standard and corporate identity. Sign-in with OpenID Connect, or with your company's identity provider. INTENTIA+ never receives or keeps the password you sign in with.
- A second factor to decide. Approving, rejecting or amending asks for a code from an app on top of the password, once per session.
- A machine never decides. It is a rule of the permission engine and of the database, whatever role the machine carries.
Connected machines, under governance
- Everything starts switched off. Assistants, programs and agents only come in once a person who administers the workspace switches them on.
- Autonomy written by a person. What each machine may do without asking, with approval or never. Before acting, INTENTIA+ answers yes, with approval, or no, with its reason and without AI. Each approval is good for one action only.
- Keys that expire. Service and agent keys are for a single workspace and always expire.
- Pause and cut off. «Pause everything» stops the connections without removing anything; «Turn off and cut everything off» revokes it all, with its reason.
- Notifications without content. When INTENTIA+ notifies other systems it sends a signed message, never the title or the text of anything.
A trail that can be checked
- Chained and signed. The Audit log, Activity and each decision's history are chained entry by entry with their fingerprint, and signed every night.
- Verify the trail. In Settings → Audit log, it walks the chain and says whether it is intact.
- It does not change. Not from the screens, not from the server, not by writing in the database.
- Every version with its fingerprint. Facts, decisions and documents keep a code that changes if a single character changes, checked every time they are read.
Your data is yours
- Your AI keys live in a secrets vault. The product keeps only a reference.
- What you delete is deleted, from INTENTIA+'s copies too. Only the record that it existed remains, with its fingerprint.
- Take it with you whenever you want. Settings → Export gives you what is yours in a file that opens in any browser, without INTENTIA+.
- What is learned carries no content. What AICA learns about each person is only for them, to switch off and delete whenever they want.
EU AI regulation
- INTENTIA+ complies with the AI Act. And it gives your organization what that regulation asks of those who use AI in their work.
- A record of what the AI did. What it was asked, what it read, which rules applied and what it proposed, in every turn.
- Real human oversight. What matters is approved, rejected or amended by a person, with their authority and their reasoning.
- Transparency about generated content. Every text an AI wrote says so, on screen and in whatever is downloaded, copied or exported.
The AI Act, article by article →
Frequently asked questions
Does INTENTIA+ train models on my data?
No. INTENTIA+ does not train models. Your text goes to your AI provider under your account and your terms, and AICA only learns from usage facts, with no text.
Where is the data stored?
On servers in Europe.
Can an AI or an agent approve anything?
No. Approving, rejecting or amending is always done by a person, with a second factor. The permission engine and the database enforce it, not the screen.
What happens if my AI key fails or reaches the spending limit I set?
The turn stops and you are told. INTENTIA+ has no fallback key.
How do I know the record has not been altered?
With «Verify the trail», in Settings → Audit log: it walks the signed chain and says whether it is intact.
Can I take my data with me?
Yes, whenever you want, from Settings → Export, in a file that opens without INTENTIA+.
Does INTENTIA+ see my passwords?
No. Sign-in uses OpenID Connect or your company's identity provider, and INTENTIA+ never receives or keeps the password you sign in with.
Need more detail for your security review? Write to us →
See how the product works →