Swiss AI Studio
Control interface for models, agents, knowledge, tools, access, status, ongoing monitoring and approvals.
View details
A modular control and operations layer for business AI.
The Swiss AI Platform is designed as the control layer behind our products and solutions: Studio controls, Agents execute and Factory assigns models. Execution runs on compute nodes – locally, in the cloud or a mix of both.
In brief: The platform is meant to separate three areas. These are the business application, control and execution with Studio and Agents, and the model layer with Factory and compute nodes. This separation can make switching easier. Testing and adjustments are still needed. We state the maturity of each building block. Studio, Agents and parts of Factory run in our own development and operating environment and are continuously extended. Further operating models, multiple nodes and private and local installations will follow step by step.
Architecture
Eight layers working together.
From application to operating model, each layer has a defined task. The shared data core (Shared Business Core) can provide customers, projects, documents and permissions to several applications.
Layer 1Applications
MABEX, Call Agent, Easy Factura, Meeting Intelligence, Offertwerk and custom client solutions
Layer 2Industry modules
Real estate, trades & building services, accounting & fiduciary, healthcare, manufacturing and finance
Layer 3Shared data core (Shared Business Core)
Customers, companies, people, projects, documents, users, roles, permissions and master data
Layer 4Swiss AI Studio
Control, configuration, oversight and monitoring
Layer 5Swiss AI Agents
Agents, tasks, tools, triggers, orchestration and approvals
Layer 6Swiss AI Factory
Access to all models, model selection per task, capability testing, load balancing, failover and quality checks
Layer 7AI nodes & models
Node 01, Node 02, Node N, local models, specialised models, external models and RAG / knowledge systems
Layer 8Operating models
On-premise, Swiss private cloud, controlled cloud and hybrid
Building blocks
Studio, Agents, Factory.
Swiss AI Agents
Execution layer: agents take on tasks, use approved tools, start workflows and request approvals.
View detailsSwiss AI Factory
Model layer for assignment, unified model access, load balancing, fallback models and quality checks.
View details
Technical platform and operation
Functions and intended benefits.
| Feature | What it means | Benefit |
|---|---|---|
| Expansion via compute nodes | The architecture is designed for additional compute nodes. | More computing power may require additional components and testing |
| Decoupled models | The application layer is separate from the model layer. | Switching models can be made easier; testing is still needed |
| Model assignment | Tasks are to be assigned by capability, data protection, speed, cost and availability. | Model selection by task |
| Load balancing | Several compute nodes and models can share the load. | Expansion based on verified need |
| Decoupled application layer | Models and infrastructure are largely decoupled from the application. | Switching models should not mean reinventing the business application |
| Operating models | On your own servers, Swiss private cloud, controlled cloud or mixed operation. | Choice based on data type and requirements |
Operating models
Local, cloud or hybrid – not dogmatic.
Sensitive documents locally, web research with an external model, transcription in a Swiss private cloud: the platform allows the right operating model for each task. Swiss AI Factory applies the data protection setting for each task, and Studio makes it visible.
Detail pages: Private AI, On-premise & hybrid, AI infrastructure.
Not public
What we deliberately do not show.
Showing expertise without revealing trade secrets (Blueprint §16).
- Internal IP addresses, ports, host names and service names
- Routing rules, weightings and algorithms of the Factory
- Security rules, firewall rules, failover details
- Specific model versions and internal version states
- Screenshots with client data, secrets, credentials
Frequently asked questions
Frequently asked questions about the Swiss AI Platform
What is the difference between Studio, Agents and Factory?
Studio controls and monitors, Agents execute tasks, Factory selects and distributes the models. Three layers, three responsibilities.
Do I need the whole platform for one product?
No. Each product runs on its own on the platform; you only see the application. The platform pays off as soon as several modules, models or operating models come together.
Can I integrate my own models?
In principle, yes: local, specialised and external models are integrated via Swiss AI Factory and assigned by capability. Whether a specific model is supported today is something we clarify in the project.
Do you show internal details?
Publicly, we show the principle, layers and data flow at a high level. Internal topology, routing rules, security rules and version states stay internal.
Is it already in production?
We state the maturity of each building block openly. Phrases such as “is designed to” describe the architecture, not a client deployment. Studio, Agents and parts of Factory run in our own development and operating environment and are continuously extended. Further operating models, multiple nodes and private and local installations will follow step by step.
Next step
For AI with requirements for data, scaling and model choice.
No-obligation introductory call about architecture, data, models and operation.