Based on a conversation between Amy Edmondson, Caspar Herzberg, Robert Newland and Lauri Paloheimo

Organizations today face a fundamental choice: treat AI as a technical upgrade or embrace it as a catalyst for systemic learning. As Amy Edmondson, Caspar Herzberg, Robert Newland and Lauri Paloheimo discuss, the path to meaningful transformation isn’t about adopting the next tool, it’s about building the cultural and psychological foundations that allow people to work and learn differently.

AI as a Catalyst, Not a Cure

Caspar Herzberg, CEO of Aveva, highlights a paradox in industrial AI adoption: while many sectors like oil and gas have been quietly using AI for decades, overall investment in AI remains marginal. This isn’t a tech problem, it’s a mindset one. Fear of failure, hierarchy-induced inertia, and data silos are bigger barriers than algorithms.

Amy Edmondson, a leading scholar of psychological safety, reframes AI implementation not as a top-down deployment, but as an organizational adaptation. In this new reality, companies become complex adaptive systems; interdependent, unpredictable, and fast-moving. Change is not managed; it is lived. This makes psychological safety essential, not optional.

Psychological Safety in the Age of GenAI

Psychological safety, the belief that one can speak up and take interpersonal risks without fear is no longer a soft skill. It is the ground floor of effective AI adoption. Edmondson emphasizes that AI doesn’t just change processes; it changes what it means to work, learn, and lead. Workers must be able to admit uncertainty and experiment without fear of judgment or punishment.

Lauri Paloheimo, co-founder of Pandatron, adds that psychological safety can now be scaled with technology. Through systemic AI coaching, people often experience being truly listened to for the first time, especially those who haven’t had access to human coaching. This AI-enabled space allows employees to reflect on fears, admit resistance, and shift from a contracted, defensive posture to an expansive, learning-oriented one.

Respecting Resistance, Redefining Failure

Both Herzberg and Paloheimo argue that resistance to change is not stupidity, it’s strategy. People resist for reasons that protect something they value, whether it’s job security or dignity. Effective transformation respects this resistance, inquires into its roots, and then builds from there.

Edmondson deepens this by distinguishing between types of failure. Intelligent failures, those made in new territory, in pursuit of learning, are necessary. What must be avoided are basic failures (avoidable mistakes) and complex ones (system breakdowns). Leaders must help their people tell the difference, and reward learning rather than punishing setbacks.

Bottom-Up Change Beats Top-Down Decree

The era of transformation as an executive mandate is fading. Herzberg notes that effective AI implementation happens where decision-making is delegated and supported by useful tools. Top-down AI programs often falter; it’s the bottom-up experimentation that drives real results. Paloheimo reinforces this: when workers engage with systemic AI coaching, they start coaching each other. Mimicry spreads. Change becomes cultural.

In sum, building AI-ready organizations is all about trust. It’s about leaders showing vulnerability, systems that tolerate intelligent failure, and spaces – human or AI – where people can speak and grow without fear. Organizations that master this will not only adopt AI. They will evolve with it.