Security and data sovereignty
Your data stays yours. Here is how.
Private AI is only worth having if you can say where your data is, who can reach it and what happens when you stop. These are the rules we work to, and what we put in writing before anything is built.
Principles
Six rules we work to
Data stays inside your perimeter
Where data must not leave your organisation, the AI runs inside your environment. Any exception is named and agreed in writing first.
Least access, read-only first
We ask for the narrowest access that does the job, and start read-only. Every access we hold is listed and can be withdrawn by you.
You hold the keys and the logs
Credentials, permissions and activity records sit with you, under the access rules you already manage.
Your data trains nobody else
What you hold is used to answer your questions and run your work. It is not used to improve a system for anyone else.
People approve what matters
Automation prepares, routes and reminds. Decisions with money, customers or risk attached stay with a named person on your side.
You can leave with everything
You own what we build for you. Documentation and handover are part of the work, so the system runs without us.
Where it runs
Three ways to deploy. You choose per workload.
The choice is made for each kind of work, written into the plan, and can change later.
On your own servers
For data that must never leave the building. You control the hardware, the network and the door.
In your private cloud
In an account and region you choose and control. Suits organisations that already run their systems this way.
Mixed
Sensitive work runs privately. Everything else runs wherever it is most economical. The line between them is drawn by you.
A low-risk start
Four steps, each one small. You decide after every step.
Free Private AI check
A few plain questions show how exposed your data is today and whether private hosting makes financial sense. Nothing to install.
Process Review
One process, mapped with the people who run it. You receive a written recommendation, including what data is involved and where it should live.
Pilot on one workload
One use, inside your environment, measured against the success measure agreed at the start.
Review, then widen
You see the result and decide whether to extend it, change it or stop. Nothing rolls forward by default.
Governance
What is agreed in writing before any build
- Data map
- Which data is used, where it lives and where it may not go.
- Access list
- Who can reach what, at which level, and how access is withdrawn.
- Change approval
- Who on your side approves each stage before the next begins.
- When something goes wrong
- Who is told, how quickly, and who decides what happens next.
- Exit
- What you keep, what is handed over and what is deleted when the work ends.
Plainly
What we do not claim
No certificates to show yet
Neoflux holds no third-party security certification today. We work under your security policies and sign your confidentiality and data-processing terms.
No published architecture
This page describes principles and outcomes. The technical detail is shared on a call, under a confidentiality agreement, so it stays useful to you and not to others.
Questions
Security questions, answered plainly.
Is our data used to train AI models?
No. With private AI, the model runs inside your environment and your data is used only to do your work. It is not used to improve a system for anyone else.
Who at Neoflux can see our data?
Only the people named in the written plan, at the access level you approve, starting read-only. Every access is listed and you can withdraw it.
Can you work under our security review and policies?
Yes. We work under your policies, complete your security questionnaire and sign your confidentiality and data-processing terms before any access is given.
What happens if something goes wrong?
The plan names who is told, how quickly and who decides the next step. Because you approve each stage, a problem is caught at a stage boundary, not after a full roll-out.
Do you hold security certifications?
Not today. We say so plainly and work under your controls instead. If certification matters for your purchase, tell us which one and we will tell you where we stand.
What happens to our data when the work ends?
The exit terms are agreed before the build: what you keep, what is handed over and what is deleted. You own what we build for you.
Bring your security questions to Neoflux leadership.
Thirty minutes, directly with the people who would lead the work. We can sign a confidentiality agreement first.