Swiss CIOs trust their AI agents, but the evidence remains patchy

29 October 2026 · Cyber security

What this is about

A survey by Red Hat of 1,400 CIOs from 14 countries, published on 27 October 2026, reveals a striking pattern for Switzerland. Red Hat does not state the number of respondents in Switzerland. All the Swiss CIOs surveyed say they run autonomous AI agents in production; 95 per cent consider these to be secure and compliant. According to a footnote, this figure is made up of 46 per cent ‘very confident’ and 49 per cent ‘fairly confident, but with some gaps’.

Why this matters for the board and executive management

The detailed figures measure different things and cannot be offset directly against the level of confidence; they do, however, paint a more sober picture. 38 per cent of respondents say they have full and continuous visibility of their open-source supply chain. 45 per cent know their data flows precisely. 49 per cent assess their AI providers during ongoing operations using measurable criteria. The strongest signal comes from confidence itself: more than half of those who are confident, 49 of 95 percentage points, admit to gaps of their own. Both of these touch on overall management and oversight under Art. 716a OR. The board must be able to rely on executive management being able to demonstrate this transparency. The IT leadership’s self-assessment is not sufficient for this.

Questions for executive management

  1. For the AI agents we use in production, can we demonstrate where their data is stored and processed, including any transfers abroad?
  2. By what criteria and at what intervals do we reassess the risks of our AI providers after onboarding?

Sources

Sources