Patrimoine
Your data stays yours.
Your data stays where it is, every access is logged, personal data is pseudonymised.
/securityA platform operated where your data lives. Sovereign if you want it to be.
Patrimoine installs in your cloud account, in your name, and operates your data where it already lives, even spread across several providers. No migration required. It does spot optimisation opportunities: duplicates, costs, residency, and proposes them to you. And if you want a fully European platform, that is a native Patrimoine option.
| Hosting | Data | Volume | Country |
|---|---|---|---|
| On-premise server · Lille | POS, payroll | 1.2 TB | FR |
| GCP · europe-west9 | sales, stock, marts | 640 GB | FR |
| AWS · eu-west-3 | e-commerce | 210 GB | FR |
| Azure · France Central | CRM | 95 GB | FR |
Example from the Maison Verdier demo, a fictional retailer.
An on-premise server, a GCP project, an AWS account and an Azure one: Patrimoine operates them together. The residency of every dataset is inventoried continuously, 24/7.
Hosting with a European provider, AI models under an “EU strict” policy: it is all there, and you choose.
The Security and Costs agents watch what costs, what drifts and what sits idle. Every opportunity is costed and proposed; migrating remains your decision.
| Store | Weekly rev. | Margin | Salaries |
|---|---|---|---|
| Roubaix | €200k | 25.2% | |
| Other stores |
Example from the Maison Verdier demo, a fictional retailer.
Everyone sees what they should. Agents included.
Identities come from your SSO. Access rules are code: versioned, and enforced in the database, the views, the API and the agents’ context. An agent never has more rights than the person it works for.
A store manager sees their stores. Finance sees margin. Nobody sees staff data except HR.
Before granting access, you see exactly what the person will see, masked columns included.
Elevated access has an end date. At regular intervals, the Security agent proposes removing what nobody uses.
Who asked for what, when, and why it was refused: the proof is there for your audit.
Personal data is spotted, then pseudonymised.
Around the clock, the agent profiles every new column as soon as it lands and tags personal data. It stays pseudonymised for anyone who doesn’t need it, agents included: no agent or AI assistant receives it in clear without an explicit right.
- A “personal” tag on the column, visible in the catalogue, enforced by the access policy.
- Pseudonymised by your rights: the same preview shows the real address to customer service and a masked ID to everyone else.
- GDPR register kept up to date by the Security agent, with the residency and retention period of every piece of personal data.
Example from the Maison Verdier demo, a fictional retailer.
Example from the Maison Verdier demo, a fictional retailer.
Vulnerabilities are handled around the clock. The report becomes a fix.
Whether it comes from the platform’s continuous scan, a tool like Vanta or a CVE notice, a vulnerability does not sit in a list. The Security agent assesses it, prepares and tests the fix in the sandbox, applies it only within the rules you have signed, and sends the evidence back to your compliance tool.
The platform’s continuous scan, alerts from your compliance tool such as Vanta or Drata, CVE notice, gap spotted in the cloud: the Security agent receives them in one place.
It assesses the exposure, prepares the fix in the sandbox and opens a PR with the evidence: Postgres 18.2 → 18.3, bucket closed, key rotated.
You sign in advance what may go through alone, such as a critical fix ready in four minutes. A structural change always waits for your signature.
Every fix is logged, and the evidence goes back to your compliance tool.
