Governance
AI Activation & Governance
From AI curiosity to AI confidence, responsible adoption that actually ships.
The problem you are facing
Your board is asking for an AI strategy. Your teams are experimenting individually, without a common framework, without risk assessment. Regulators are moving faster than your governance: EU AI Act, Swiss nDSG, FINMA directives.
You have seen the impressive demos. You have launched one or two POCs. But between experimentation and production deployment, there is a chasm. Nobody in the organisation knows how to classify data for AI, assess the risks of a model, or estimate the real cost of LLMs in production.
The risk is not failing to adopt AI. It is adopting it poorly: without governance, without a risk policy, without regulatory alignment, and with costs spiralling out of control.
My approach
I combine deep technical expertise in AI/ML with hands-on governance experience in regulated sectors (Geneva private banks, FINMA, APRA). My approach is built on a 7-axis AI maturity model, aligned with ISO 42001 and the EU AI Act, while remaining pragmatic and adapted to the Swiss context.
AI sovereignty: a deliberate choice
AI sovereignty does not mean rejecting the cloud or hyperscalers. It is a deliberate choice: what runs where, based on risk, regulation and strategic priorities. Some workloads are perfectly suited to public cloud. Others demand total control. I help you articulate the three zones:
Speed
Public cloud
Azure OpenAI, Anthropic, Google. Fast, integrated, enterprise support. Suited for non-sensitive workloads.
Balance
Private cloud
Open-source models (Mistral, Llama) on dedicated infrastructure. Data control, predictable cost, portability.
Control
Sovereign on-premise
Local LLMs for sensitive data (banking secrecy, patient data). Maximum compliance, minimal latency.
AI FinOps and data protection
AI FinOps: controlling costs
CFOs are asking: "What will AI cost at scale?" Without governance, LLM costs spiral. I help you put in place:
- ✓ Predictive cost modelling per use case
- ✓ Token consumption monitoring
- ✓ Model optimisation (right-sizing, caching, routing)
- ✓ ROI framework for AI investments
Data protection for AI
You cannot feed your LLMs with unprotected sensitive data. My 4-step approach lets you fuel AI safely:
1. Discover
Map sensitive data
2. Tokenise
Replace with secure tokens
3. Augment
Generate synthetic data
4. Apply
Feed RAG safely
Certifications and expertise
In addition: MIT AI Programme (2024), Cambridge CTO Programme (2018), AWS Certified (2017). 30+ years of experience in architecture and IT governance in regulated sectors (FINMA, APRA, nDSG, GDPR).
Engagement formats
Entry point
AI maturity diagnostic
Complete assessment across 7 axes, maturity scoring, identification of regulatory gaps, 90-day action plan. You leave with a costed roadmap.
- 5 days
- Deliverable: maturity report + roadmap
- From CHF 8,500
Build
AI governance framework
Suite of policies, RACI, risk methodology, use case assessment matrix. 2 awareness workshops included.
- 10-15 days over 2-3 months
- CHF 18,000 - 25,000
- Aligned with ISO 42001, EU AI Act, nDSG
Certification
ISO 42001 preparation
Complete AIMS documentation, internal audit, management review, training. Objective: ready for certification.
- 15-20 days over 4-6 months
- CHF 28,000 - 40,000
- Objective: ISO 42001 certification
Activation
AI activation sprint
One concrete use case from idea to working pilot, with governance built in from the start.
- 8-10 days
- CHF 14,000 - 18,000
- Includes: risk assessment + FinOps
Full journey (diagnostic + framework + activation + certification): CHF 55,000 - 75,000 over 6-9 months.
Who is it for?
Technical directors (CTO/CIO)
Who need to structure AI adoption beyond POCs and present a credible strategy to the board. Need a maturity model and roadmap.
Compliance officers / CISO
Who need to assess AI risks, classify data, and put in place a governance framework before the regulator demands it (FINMA, nDSG, EU AI Act).
SME CEOs (50-500 employees)
Who see AI transforming their sector and want to act in a structured way. Need a framework without hiring a specialised team.
Regulated sectors
Banking, insurance, healthcare, public administration. Where compliance is not optional and every use of AI must be documented and traceable.
Field experience
I have designed and deployed AI governance frameworks for Swiss financial institutions regulated by FINMA. I modernised the data protection and anonymisation platform of a Geneva private bank, then coordinated the data migration as part of a major acquisition. I also created a Strategic Data Protection Council with the CISO and DPO.
I also designed a 7-axis AI maturity model for a Geneva private bank, covering governance, cybersecurity, regulatory compliance, infrastructure, data quality, skills and application portfolio.
Book a conversation
A 30-minute conversation, not a sales pitch. We will discuss your situation and I will tell you honestly whether I can help.