Data
Where can data, prompts, outputs, logs and backups go?
SovereignEU Trust Centre
Sovereignty you can inspect.
Understand where data can go, who operates the platform, which components are involved, how AI is controlled and how your organisation can remain portable.
Sovereignty matrix
Six questions every deployment should answer.
Each claim should lead to a dated policy, architecture document, test result, licence record or deployment-specific disclosure.
Operational
Can your organisation keep operating without a mandatory vendor cloud?
Technology
Can components be inspected, integrated, adapted and replaced?
Jurisdiction
Which entities, staff and subprocessors can access the service?
Knowledge
Can your teams understand, operate and improve what was delivered?
Exit
Can data, policies, models and workflows move in usable formats?
Operational boundary
Know where every path leads.
A deployment diagram should show data, prompts, outputs, logs, backups, administrative access, updates and support paths.
Disclosure by deployment option
- Operators and administrative access
- Components, licences and model provenance
- External services, telemetry and subprocessors
- Backup, updates and vulnerability handling
- Export, migration and operational handover
AI governance
Evidence, not a badge.
Inventories, evaluations, version history, policies and events can support organisational governance. Framework alignment is not a certification or legal guarantee.
Explore AI Guard
Exit readiness
Portability is an engineered capability.
Data exports, configuration transfer, model and workflow portability, migration assistance and handover should be designed and tested.
Explore servicesA practical first step
Bring your technical, security and procurement requirements.
Review the evidence and boundaries that your organisation will need before deployment.