Evidence,
not assurances.
“We have it under control” is not an answer an auditor accepts. Here is what we can show: how AI actions are approved and logged, where your data lives, and who else processes it.
Three questions,
three answers.
AI that acts inside your network raises the same three questions from every auditor, regulator and board. Who approved this action? Where does the data live? Who else touches it? We answer each one with evidence you can check.
On record
Consequential changes wait for approval. Every step is logged and reversible.
In your jurisdiction
Region-pinned processing, under Swiss law at the top of the group.
On the table
Every subprocessor disclosed, by entity, country and purpose.
Every consequential action
waits for a human.
Lucy and her specialists propose, a human approves, and only then does the change run. Every change arrives as a diff, an impact simulation and a request for your approval. Bounded, reversible actions run immediately; consequential ones wait. Every step is logged and reversible, with full provenance on every action.
You set the limits: dial autonomy up or down per domain and per risk level. In Mission Control, a Level-3 engineer approves before anything touches production.
| Time | Actor | Action | Approval | Scope |
|---|---|---|---|---|
| 09:14:02 | Lucy | Root cause found: misconfigured ZTNA policy, site-DE-04 | Diagnosis only | read-only |
| 09:14:05 | Lucy | Rollback proposed: diff + impact simulation | Awaiting approval | policy |
| 09:16:40 | Admin (you) | Rollback approved | Approved | policy |
| 09:16:41 | Lucy | Rollback applied, revision 2291 | Reversible | policy |
| 09:30:00 | Calli | Evidence pack generated | Exported as PDF | report |
Exportable evidence: Calli, the specialist for reporting and evidence, builds evidence packs mapped to the controls auditors ask for, every figure sourced, as PDF, live page or scheduled send. Meet Calli
Attributable by design: through the API you approve, decline or hold change requests, self-service changes reference their audit trail, and API users are dedicated machine identities with scoped permissions. Developers & API
A jurisdiction you can point to.
Data residency inside a hyperscaler does not erase the jurisdiction of its parent company. So we show you exactly where you stand, and where the limits are.
- Swiss at the top: Open Systems is part of the Swiss Post Group, a Swiss federally owned company, so the ultimate parent sits under Swiss law.
- Region-pinned processing in your jurisdiction, with subprocessor disclosure, data-flow maps and documented residual risk.
- Fully Swiss-resident option for Swiss-banking workloads, FINMA-aware.
Your model provider is a jurisdiction question.
Not only a procurement one. Where a model runs, and who runs it, decides whose law applies to your prompts, your telemetry and your answers. So we keep the model a choice, not a lock-in.
No single-LLM lock-in
The best model per task, inside a compound system that is grounded on real telemetry, with citations.
Inference in your region
Inference can be pinned to your jurisdiction instead of following a vendor’s default cloud.
Raw data never trains for others
Lucy is scoped strictly to your tenant. Cross-customer learning happens as patterns only, and only where compliance allows.
Every entity that may process data on our behalf is listed in our subprocessor statement. Why this layer deserves its own audit: AI Sovereignty: The Fourth Risk Layer You Have Not Audited
Every subprocessor on the table.
Our subprocessor statement lists every entity that may process data on our behalf, in three groups, each with its country and its purpose.
Affiliated entities
Open Systems companies in Switzerland, the United States, Germany and India.
→PlatformThird parties for the platform
Infrastructure and services behind Mission Control, such as Microsoft Azure.
→Per serviceOnly if you use the service
Specialists for single services, for example a cloud region you choose yourself.
→Audited every year, not once.
27001, 27017, 27018
Certified information security management, including the cloud-specific controls and the protection of personal data in the cloud.
Type 2 report
Assessed annually by an independent auditor.
Reports on request
Details of our ISO and SOC certifications are shared on request.
→Mapped to what you report against.
- NIS2 and DORA readiness: map controls to the frameworks you report against and close the gaps with expert support.
- GDPR and FADP: data sovereignty in line with European and Swiss data protection law.
- One auditable timeline: what happened and who acted, a natural audit trail.
Audit Readiness: How Open Systems Takes the Pain Out of Compliance
Questions, answered.
Can the AI change my production network on its own?
Can I export the evidence?
Where is my data processed?
Which certifications do you hold?
Ask for the evidence,
not the brochure.
Details of our certifications and audit reports are shared through our Trust Center.
Visit the Trust Center →