Services · Governance
AI Governance & Training.
AI in a company only works when the team uses it and knows what for. This service gives your organisation a lived frame: an AI usage policy that opens paths instead of closing them, documented AI Act readiness and training that lets your team use AI safely.
When governance becomes relevant
Shadow AI needs an official path
AI is used in-house, but without a clear frame. Shadow AI shows that your people already see the value. The policy turns quiet practice into a managed one.
Documenting AI Act duties properly
Documentation, transparency and competence duties can be built into the architecture path. Considered early, they simply run alongside the build.
The team should use AI safely and productively
The AI Act requires sufficient AI competence among employees. The training fulfils this duty and at the same time shows your team where it can use AI day to day.
What you get
AI usage policy
A policy tailored to your organisation: what is allowed with which data, in a form that is lived day to day.
AI Act readiness dossier
Risk classification of your AI applications, the necessary evidence and documentation, provable towards supervision and audit.
Role-based training
Training for management, business units and IT, using your own use cases. Fulfils the AI Act competence duty.
AI officer (on request)
Setting up the role and supporting its first steps, if you establish a responsible position for AI.
The process
- 01
Taking stock
Which AI is in use, which duties apply, where the organisation stands: the basis for policy and training.
- 02
Policy & dossier
The AI usage policy and the AI Act readiness dossier are drafted and aligned with you.
- 03
Training & anchoring
Role-based training of employees and the hand-over, so governance is lived.
Price & scope
from CHF 3’000
Policy, readiness dossier and training are agreed individually or as a package, depending on scope. Half-day training formats pro rata.
What follows from it
You can prove the responsible use of AI: towards supervision, management and employees. Governance stays a standing task after that; on request, Souverana supports it on an ongoing basis.
References
Workshops and keynotes on sovereign AI for Swiss companies: CH Open Source AI Conference, Swiss AI Conference, SWICO keynote. Hands-on, not slides.
Teaching
Workshops at the CH Open Source AI Conference and the Swiss AI Conference, SWICO keynote on AI for Swiss SMEs, panel at the Swiss AI Impact Forum.
Platform
Architect of the Swiss AI Hub: 35 integrated containers, air-gapped on a single GPU or on Swiss cloud. 15 projects, 12 companies, 2 government tenders.
Research
ConceptFormer: published at The Web Conference (WWW ’26); +348% factual accuracy when grounding language models in knowledge graphs.
Education
MSc in Computer Science, University of Zurich (summa cum laude, AI major, master’s thesis grade 6.0).
Entrepreneur
Co-founder of PolygonSoftware (2019–2023): 62 projects for 25 clients across five industries in four years, incl. a computer-vision system for Swiss Post.
Network
Known participant in the Netzwerk Digitale Selbstbestimmung (netzwerksds.ch), via bbv.
Frequently asked questions
- Does the EU AI Act even apply to us?
- In many cases yes. It also affects Swiss organisations as soon as they use AI or operate in the EU market. Depending on the application, documentation, transparency and competence duties apply; employees must receive training. The stocktake clarifies exactly how you are affected.
- Is a template from the internet enough as an AI policy?
- As a starting point perhaps, as a lived policy no. A policy only works when it fits your data, your industry and your tools and is understandable in daily work. That is exactly the work here.
- What does the training duty mean for us?
- The EU AI Act requires employees to have sufficient AI competence. The role-based training fulfils this duty and is at the same time the practical part: afterwards your team knows how to use AI safely.
- What experience is behind the training?
- Workshops and keynotes on AI for Swiss organisations: at the CH Open Source AI Conference, the Swiss AI Conference and as a SWICO keynote. Teaching comes from practice.