We build your company's best employee.
Working around the clock, learning every day.
AI agents that take over your most expensive processes: connected to your systems, carrying your whole company's knowledge, and learning from every correction your team makes. When your best specialist retires one day, their knowledge stays in the house. Run on your private Swiss workspace, optionally on models that never leave Switzerland. And your team approves every action.
According to contract no. LV-2025-117, late delivery triggers:
- –Contractual penalty of 0.5% per started week, capped at 5%
- –Claims for further damages remain reserved
- –Warranty: 24 months from delivery
Sample answer from the Acteon demo, based on fictional documents.
Your digital employee in action.
The first example: quoting. Read the enquiry, check it against your pricing and article data, present a draft for approval. In 80 seconds.
(Video in German)
Your employees already use AI. The question is where your data ends up.
Public AI tools are uncontrollable
Whatever employees type into ChatGPT is processed on US servers. You cannot see it, steer it or get it back.
Bans do not work
Ban AI and you get shadow usage on private devices. The risk stays, the productivity gain is lost.
Building it yourself is slow and expensive
An internal AI project ties up capacity for months, costs six figures and is outdated before it ships.
What else you get: an AI workspace your company owns
The quoting agent runs on your private AI workspace: a dedicated instance under your address, with your logo and your rules. Familiar to use, fully under your control.
The best models, always current
Claude, GPT and Swiss models such as Apertus (ETH/EPFL) in one interface. New model generations included.
Company knowledge with sources
The AI answers questions about your contracts, projects and documents and backs every answer with the source file.
Create documents directly
Word, PowerPoint, Excel and PDF straight from the chat, generated on your Swiss instance, ready to download.
Dictation and minutes
Meeting recordings and dictations are transcribed on your instance and turned into structured minutes.
Role-based access
Management, departments, administration: who sees what is enforced server-side. We demonstrate it live.
Research with sources
The research assistant searches the web repeatedly, checks contradictions and delivers structured, sourced reports.
You decide per task where computation happens
Two tiers, one click. Storage is always in Switzerland.
Maximum quality
EU- Where does it compute?
- Processing in EU data centres (GDPR area)
- Best for
- Complex analysis, demanding drafts, research
Maximum sovereignty
CH- Where does it compute?
- Processing in Geneva, never leaves Switzerland
- Best for
- Highly sensitive content and documents
Your employees pick the tier right in the chat, per task, with one click. Documents, knowledge base and chat history live exclusively on your dedicated instance in Geneva.
We tell you exactly where your data is
No vague promises: for every component we disclose where processing happens, contractually guaranteed.
Why we do not promise 'everything runs in Switzerland'
The world's most capable AI models do not run in Switzerland today. Promises like 'everything runs in Zurich' rarely survive scrutiny. We do it differently: frontier models with documented EU processing, Swiss models for your most sensitive work, and every route disclosed in writing. Verify us.
How your documents stay private
The AI works with your files without them ever leaving your control.
Only the relevant excerpt, only briefly
For each answer, only the needed passage is sent to the model, processed and immediately discarded. Your full document corpus stays on your Swiss instance.
Access only where authorized
For each person the AI sees only the documents they are cleared for. Roles and groups are enforced server-side, not merely hidden.
No training, deletable anytime
Your content never trains third-party models, contractually guaranteed. On request we export everything and delete it irreversibly. A data-processing agreement under revFADP and GDPR governs the details.
Acteon next to the standard tools
| ChatGPT | Microsoft Copilot | Acteon | |
|---|---|---|---|
| Where is your data? | USA | Shared cloud, EU | Switzerland, own instance |
| Training on your data | Possible depending on plan | No | Contractually excluded |
| Answers with sources from your documents | No | Limited | Yes, with citations |
| Swiss model tier for sensitive work | No | No | Yes, Apertus and Qwen in Geneva |
| Onboarding and training | None | None | 4 sessions included |
| Documents, dictation and research | In the US cloud | In the Microsoft cloud | On your Swiss instance |
From demo to live operation
- 1
Demo
45 minutes, live, on real examples. Bring a typical enquiry.
- 2
Design partnership
We build your first workflow agent on real data: first win in week 1, go-live in 6 to 8 weeks. CHF 5,000, fully credited against the setup if you continue.
- 3
Operation
Updates, support and reviews run in the background. You focus on your business.
Frequent questions
Where exactly do the AI models run?+
The Swiss tier (Apertus, Qwen) computes in Geneva. The quality tier (Claude) computes via AWS Bedrock in EU data centres (GDPR area); GPT is optionally available via Microsoft Azure EU. Storage is always in Switzerland. You receive the full processor list contractually.
Is our data used to train AI models?+
No, on no tier. This is contractually excluded, for us and for the model providers we use.
What happens to our data when the contract ends?+
You receive a complete export in open formats. We then verifiably delete everything within 30 days.
Can employees sign in with their Microsoft or Google accounts?+
Yes. Single sign-on with Microsoft 365 is available as an add-on, including automatic group and role sync. Sign-in with Google Workspace is supported as well.
How fast are we productive?+
The design partnership gets your first workflow into production in 6 to 8 weeks. The full setup takes four weeks from kickoff, training included.
What does model usage cost?+
Nothing extra. All model usage is included in the monthly fee, with no token billing.
How long are we committed?+
Minimum term of 12 months from go-live, then cancellable with 3 months notice. The design partnership is independent and ends after 6 to 8 weeks.
Who can access our content?+
Only the users you define, according to your groups and roles. Acteon accesses content only for troubleshooting, after informing you first.
See it on your own examples
45 minutes, live. We show the platform on real tasks from your day-to-day.