Model Drift: When an AI System No Longer Matches the Business Case
An AI system can change without anyone changing it. Why model drift is a governance duty, and how the board should operationalise it.
Information for boards and executive teams
Perspectives on AI, governance, cybersecurity and digitalisation for boards and executive teams.
The Duty-of-Care Trap: When the Board Must Supervise What It Cannot Inspect
AI systems often cannot be fully audited. Why the board's supervisory duty must become proxy oversight, not the illusion of transparency.
Autonomous AI in Operations: The Accountability Questions the Board Must Resolve
As AI systems adapt in operation, accountability becomes a governance question: which controls must the board demand and document?
Shadow AI: Why the Board Needs to Know This Risk
One in five organisations had a shadow AI security incident in 2025. Three levers the board can use to bring this risk under control.
Technology Competence in the Executive Team: Why Efficiency Often Fails Without It
36% of Swiss board members name efficiency as their biggest opportunity. Without technology competence in the executive team, it stays a promise.
The Dual Role of CEO and Chair: Limits and Risks
Suzanne Thoma leads Sulzer as both Chair and CEO. When this dual role works, and where its structural risk actually sits.
The Danger of Automation Bias: Is Human-in-the-Loop an Illusion?
A human watching the screen does not by itself prevent errors. What the EDPS reveals about effective oversight of automated decision-making.
AI Hallucinates. Does the Board Notice?
Hallucinations are not a teething problem of AI, they are the architecture. Six measures the board needs to contain this governance risk.
Data Sovereignty: Between Illusion and Real Control
A European cloud is a procurement decision, not a sovereignty strategy. Real control is architecture, technically enforced, and it is a board question.
Why AI projects fail and what it says about the organisation
AI pilots often succeed while the rollout fails. The reasons are rarely technical; they are organisational: data ownership, accountability and process design.