AI learning journey London

Where will AI lead us?

Facilitator: Carlo van de Weijer

AI is a result of exponential development that has been underway for a very long time. The challenge lies in interpreting the distinction between human and machine flexibly, whilst remaining acutely aware of the risks. As organisations place more trust in AI across more functions, the risks grow exponentially. We are at a critical moment where early investors are beginning to see productivity gains, but this comes with a verification cost. Everyone has ideas about what should be implemented with AI, creating execution shortages.

The framework is simple: prompt, work in the middle, check and own. This requires genuine intelligence and craftsmanship. When AI takes over the middle work traditionally done by junior staff, organisations stop developing senior talent. AI must be understood as a strategy, not merely a tool. New technology fundamentally requires you to change your entire business; otherwise, you become less flexible.

The advice is to experiment sensibly with AI whilst being aware of the top risks. As a non-executive board member, the challenge is to set guidelines whilst leaving room for flexible interpretation. Rather than asking "What is our AI strategy?", the more important question is "How does AI change our strategy?

Key takeaways

• AI requires changing your entire business process, not just adding a tool
• Governance is critical. Determine who has access and who can create
• Experiment sensibly whilst managing verification costs and execution capacity
  

Strategy led AI: ownership and explicit choices

Speaker: Jorg Schalekamp

Whose responsibility is AI? The answer is clear: while ultimate ownership and execution belong to the CEO, achieving true success requires a strong partnership with Non-Executive Directors. Non-Execs play a critical role in helping the CEO successfully navigate the complex AI landscape. Together, they must ensure the organisation is explicit about where AI will deliver value, whether in supporting functions or at the core of operations. Non-Execs can provide the essential outside-in perspective needed to evaluate where the organisation currently stands, helping the CEO define value spaces, identify core processes, and effectively move from assets to execution and productivity gains.

Portfolio selection becomes a crucial collaborative question: what do you do first, and what follows? Non-Execs help the CEO move from generalist exploration to deliberate, prioritised investment, avoiding both market irrelevance and unfocused spending. This requires explicit choices about core versus supporting functions, and active management of risks such as vendor lock-in.

Here, Non-Execs serve as a vital sounding board, guiding and challenging the CEO to make explicit, well-calibrated choices about where to invest and where not to invest. This strategic oversight ensures that AI investments are not just experimental, but are steered toward generating tangible returns and preventing the market irrelevance of the organisation. Furthermore, Non-Execs can encourage the CEO to look beyond operational gains to discover new propositions, while maintaining rigorous oversight on the total costs of AI, the components of the AI strategy, and the critical need to avoid vendor lock-in.

Key takeaways

• Ask how AI changes your strategy, not what your AI strategy is
• AI ownership belongs to the CEO, championed by Non-Execs: The CEO drives execution, while Non-Execs provide the guidance and oversight necessary for the CEO to be successful.
• Collaborative strategic alignment: Non-Execs help the CEO be explicit and rigorous about where AI will deliver core organizational value.
  Guided investment choices: Non-Execs support the CEO in portfolio selection, ensuring deliberate investment choices that generate clear returns and avoid market irrelevance.
• Risk and opportunity oversight: Together, they navigate the costs, discover new propositions, and actively manage risks like vendor lock-in.
  

Build your own: from theory to practice

Facilitator: Arthur de Wilde & Mira Deutsch

We talk a lot about AI, but its actual usage remains difficult. The reality is that we use it too little. This interactive workshop bridges the gap between theory and practice. Participants build startups in groups, practice with all available AI tools, upload outputs, and present results. This hands-on approach turns understanding into capability.

Key takeaway

• Close the gap between AI discussion and actual implementation.
• Hands-on experimentation with real AI tools builds confidence, capability, and clarity about genuine possibilities and constraints.
• Avoid using AI in isolation. Run multiple teams in parallel on the same problem to generate n fold outcomes, increase diversity of ideas and bypass traditional hierarchical bottlenecks
  

Wisdom of the crowd: building the foundation

AI adoption is a strategically fundamental change process that is less about technology than many expect. Success requires a long-term perspective. Think in terms of a five-year plan. Building a strong foundation is essential, and organisations must understand that AI is not free; it requires significant investment.

For board members, the question becomes how far do you step in, to influence what the board does and tackles. The opportunity is to narrow the gap at the front end with developments, recognising that this represents a much larger change than in the past. The role is to make the CEO successful whilst bringing your own responsibilities to the table, inspiring dialogue and engagement. AI must become a permanent agenda item. It should be acceptable to discuss trying things that might fail.

The board must determine what constitutes good choices and good use cases, defining how these contribute to organisational strategy and balancing efficiency gains against necessary investments. Choosing use cases and smart partnerships is critical. Speed can be gained by using what already exists. The main question that must be answered is; what is actually our plan?
The insight is powerful. Use AI to strengthen your strategy. Rather than suddenly doing different things because of AI, ask how AI helps you do what you already do well, even better.

Key takeaways

• AI adoption requires a long-term, five-year perspective
• Make AI a permanent board agenda item
• Balance efficiency gains against necessary investments
• Use AI to strengthen existing strategy, not to chase new directions
  

Leadership in the age of AI

Facilitator: Gianpiero Petriglieri

Leadership actively emerges when circumstances challenge our adaptability. Facing an existential juncture or a gap between experience and expectations, people look for leadership to avoid becoming disoriented, distressed, and disconnected. Two paths forward: a defensive path that shrinks and a developmental path that expands. Leadership is fundamentally about experiencing and feeling, showing itself in how we behave.

The dominant model of leadership centers on position, property, attention, and influence. Its holy trinity consists of vision, strategy, and persistence. Leaders ask people to invest work today for returns tomorrow, requiring trust. However, leadership itself is a relationship where people feel safe and free, committed but not captive. Quality leadership balances hope against fear, work shown against work asked, and freedom against demand.

Leadership is essentially an expanded circle of care that shrinks as anxiety rises. Passion alone is insufficient; preparation is necessary. It is a performance of story-giving rather than story-telling, expressed in every choice and gesture. A better leader balances patience and pressure. Since learning is the new loyalty, a caring leader gives you what you need, operating as a custodian providing what is needed, a challenger questioning why you do things, and a connector sharing challenges and fostering collaboration.

Key takeaways

• Leadership emerges when we face circumstances that challenge our adaptability
• Leadership is a relationship built on care, not a position
• Balance hope with fear, work shown with work asked, freedom with demand
• For AI: ask how do I make it matter, these are human questions
  

Responsible AI: beyond the algorithm

Facilitator: Alix Rübsaam

Designing an algorithm fundamentally means translating complex operational goals into automated, rule-based decisions that carry significant, real-world consequences. While these tools are highly effective at optimizing efficiency and identifying explicit data patterns, they are fundamentally limited when applied to nuanced human environments. Within any organization, group dynamics and the qualitative ways people relate to one another matter immensely to long-term success.

A critical vulnerability of algorithmic design is that core human values, most notably integrity, cannot be measured, quantified, or coded by a machine. This limitation introduces profound, fundamental questions regarding how an organization can guarantee a truly responsible AI rollout when the technical tools themselves are inherently blind to the principles that matter most. Relying solely on data inputs can lead to severe unintended outcomes, such as automated recruitment tools replicating historical societal biases against specific demographics. True corporate responsibility requires recognizing that automated systems are logic engines, not ethical entities; therefore, maintaining an active layer of human oversight is mandatory to preserve institutional integrity.

Key takeaways

• Algorithms carry real consequences: Automated, rule-based decisions lack empathy and directly impact human lives, requiring careful ethical boundaries during their design phase.
• Integrity cannot be encoded: Core ethical principles like fairness, empathy, and integrity cannot be reduced to math or measured by an algorithmic tool.
• Human oversight is mandatory: Because technology cannot evaluate qualitative human dynamics, continuous human judgment is essential to ensure responsible and safe AI implementation.

Conversational AI: The power of voice

Facilitator: Jack Smith

Conversational AI is changing how organisations interact with their customers. One powerful application is helping people who have lost their voice through voice cloning technology. The use of voice is significant because it conveys intent to action. Voice is more personal, emotional, and memorable than text. This technology allows companies to build their own speech agents and avatars, creating more human-like interactions.

A critical consideration is embedding AI (the technical aspects) in non-technical departments of the organisation. This requires control over what is being done and enabling AI in a way that provides visibility into operations. Organisations must take full accountability for what AI produces. You are responsible for the output, which means reviewing and approving what AI generates before it reaches customers or stakeholders. Embedding AI in non-technical departments requires visibility and control; otherwise, accountability becomes impossible. This means reviewing and approving what AI generates before it reaches customers or stakeholders.

Key takeaways

• Value of AI sits in specific applications like conversational solutions to create human-like support
• Embed AI in non-technical departments with visibility and control
• Take accountability for what AI produces, review and approve

Legal and regulatory development

Facilitator: Jan-Jan Lowijs

The regulatory landscape surrounding artificial intelligence is undergoing a rapid, global evolution, presenting complex challenges for corporate governance. As organizations accelerate their AI adoption, board members must thoroughly understand the shifting legal implications to effectively guide their enterprises through implementation. A common misconception is that AI operates in a legal vacuum or that new frameworks replace existing structures. In reality, upcoming legislation like the EU AI Act serves primarily to fill specific regulatory gaps.

Every AI system remains completely bound by "regular" established laws, including privacy (GDPR), non-discrimination, antitrust, and copyright legislation. Claiming management did not know or intend for an algorithm to breach compliance is not a legally viable defense. Because AI systems leave behind an extensive, highly transparent digital footprint of their operational outcomes, compliance errors and biases are much easier for regulators to detect and scrutinize than human mistakes. Consequently, boards must proactively invest in robust compliance frameworks and ensure their legal teams receive technical training to understand how these systems function.

Key takeaways

• Regular laws apply fully: AI systems are completely subject to existing frameworks, meaning traditional privacy, consumer protection, and non-discrimination laws apply to algorithmic outcomes.
• Ignorance is no excuse: Corporate leadership cannot bypass accountability or avoid penalties by claiming they were unaware of how a deployed algorithm functioned.
• Heightened regulatory scrutiny: Because AI workflows leave precise digital footprints, compliance errors and underlying biases are highly transparent and easily targeted by supervisory authorities.
• Non-execs (non-executive directors) require both greater breadth and greater depth than executives.
• Because many non-execs hold multiple roles across different contexts, they face a wider and less generic set of dilemmas and questions — and they need to understand the implications of decisions at a deeper level.
• Where executives often look for best practices around “what’s possible” and “what’s practical,” non-execs are more focused on the underlying mechanisms, dynamics and trade-offs. For example: “At one company I ask how far to let management run before risking an inability to catch up; at another I ask how much risk and investment to allow so the company can leap forward.”