This episode features, in order of speaking, Doug Else-Jack, Fiona Revell, Mariann Sahni, Cameron Smith, Pramod Prasanth, Corrado Mazzoni and Marc Rajal.
SWISSUES participants choose what aspects of the topic they want to address and this episode cover the following.
Credentials are not experience
Credentials and experience are routinely conflated, but they’re not the same thing — and the gap matters most where credentialing culture is light (Switzerland was raised as a case in point, against markets where degrees are the norm). A useful reframe: experience isn’t a credential at all, it’s an engine — the mechanism by which learning, adaptation, and leadership happen. Translating experience into a new context (a CV line, an interview, a lateral career move) is its own separate skill.
Experience doesn’t protect against AI-driven mistakes — it may increase the risk
A central claim of the discussion: experienced people tend to overtrust AI output more than less experienced people, because they scrutinise it less. The Air France Rio–Paris crash came up as a real-world parallel — experienced pilots over-trusting an automated system. Familiarity makes a wrong answer feel right. One proposed counter-measure: deliberately withhold the polished conclusion in decision-support tools and force users to check the underlying numbers themselves, to keep the analytical muscle alive — especially for juniors who’d otherwise never build it.
The “borrowed slide deck” problem
A vivid analogy for using AI without understanding it: presenting someone else’s slide deck, and only realising on slide three that the underlying logic can’t be explained. Most people only need that experience once before changing how they use AI.
What AI still can’t capture
A recurring theme: AI has no access to organisational culture, atmosphere, or “the room” — things only available through physical presence. AI was framed as a data engine: useful for synthesis, but entirely shaped by what humans feed into it. Humans remain the validating layer, and AI may end up freeing time for more human interaction rather than less.
A live governance question
A relayed example: someone coding extensively with AI agents who admitted he can no longer fully track what those agents are doing — raising an open question about where human oversight sits once AI-assisted work outpaces a person’s ability to review it.
Passing experience on: coach, don’t tell
Strong convergence on this point — mentoring works by walking someone through a problem so they reach their own conclusion, not by handing them the answer. (A parenting parallel: telling a child what to do rarely works as well as letting them live it.) Experience only transfers if it’s translated into terms the other person can actually use — a lawyer’s experience doesn’t help a mechanic without deliberate translation.
Closing note
Despite real enthusiasm for AI’s usefulness, the overall sense was that accumulated human judgement and wisdom aren’t replaceable yet — and may never be.
Next SWISSUES Forvm: Thursday 2 July — “Hive-Mind,” on how decisions emerge in organizations without ever being formally made.
