How leaders can become AI-savvy without becoming technologists

Executive overview

Most AI adoption projects fail not because of bad technology, but because business leaders abdicate responsibility to tech experts. The real challenge is behavioral: getting people to trust, adopt, and work alongside AI tools that are imperfect from day one.

AI adoption is a leadership challenge, not an engineering exercise — and treating it as the latter is why 80% of projects fail to create value.

Why AI is different from previous technology shifts

• Past innovations outsourced physical capability; AI outsources cognition, raising existential anxiety about what remains distinctly human.
• Business leaders today expect more from AI than they expected from the internet in the 1990s — and those inflated expectations drive both hype and disappointment.
• The productivity gain from AI follows a J-curve: performance dips initially before rising, because both the AI and the humans using it need time to adapt.
• 80% of AI adoption projects fail to scale — twice the failure rate of non-AI digital transformation projects.
• Failure is rarely in buying the AI; it's in never creating value across the organization's stakeholders.

The two perspectives on AI — and why one leads nowhere

• Perspective one: AI as a cost-cutting tool to replace people and maximize efficiency. Treats AI as the end in itself.
• Perspective two: AI as an augmentation tool that amplifies what humans are already good at. Treats AI as a means to an end.
• Perspective one taken to its logical conclusion leads to eliminating humans entirely — which is not a viable long-term strategy.
• Human intelligence and artificial intelligence are apples and oranges; they are not in competition.
• Simply adopting AI does not create competitive advantage. Adopting AI to enhance what you are already strong at does.
• Most business leaders remain stuck in perspective one, framing AI as a zero-sum game.

Why business leaders must stay in the room

• AI adoption is treated as an engineering exercise, so business leaders step back and delegate to tech experts.
• Tech experts are not trained to ask the right business questions — and everything starts with a business question.
• Without the right question, no AI deployment will create value, regardless of how sophisticated the model is.
• When AI projects fail, a consistent pattern emerges: business leaders were absent from the process.
• Leaders must be AI-savvy enough to define the relevant business question, then hand off to tech experts — not the other way around.

What AI savviness actually means

• AI savviness is not about coding; it is about understanding what AI can and cannot do.
• Understanding AI's limitations reveals what is uniquely human — and those human capabilities are what the AI should be built around.
• AI-savvy leaders understand statistics and the difference between correlation and causation.
• The lab model shown in demos is never the model deployed in your business. In a business context, AI adoption becomes a risk management issue, and supervised AI remains the norm.
• Leaders do not need to track every new large language model; they need the foundational concepts so they can evaluate and integrate new tools as they emerge.
• The distinction that matters: an AI-driven company (AI leads) versus an AI-enabled company (AI serves human goals).

Building the conditions for AI adoption

• The most expensive part of AI is not buying it — it is implementing it: training people, restructuring workflows, building infrastructure.
• AI models are not optimized on day one; they improve through feedback. Job design must leave room for that feedback loop to operate.
• AI adoption is a collective, lifelong learning process — not a one-time deployment event.
• Leaders must make AI meaningful to employees: show it is a means to organizational purpose, not a replacement threat.
• Managers across departments and frontline teams all need to be brought along, not just the technology function.

What leaders apply that tech experts cannot

• Core leadership skills — communication, emotional intelligence, vision, and mission — are exactly what AI adoption requires.
• AI-savvy leaders act as narrators: they bridge tech experts and business experts, breaking down silos.
• They know how to build trust, manage change, and empower people to buy in — the behavioral work that determines whether adoption succeeds.
• Leadership today is evaluated not just on traditional competencies but on the ability to make sense of AI in the context of organizational goals.

What David De Cremer changed his mind about

• We are as responsible for AI hype as AI itself. The confidence AI projects is largely a human perception — we attribute meaning and capability beyond what the technology warrants.
• Every major innovation or crisis follows the same pattern: inflated expectations, a reality-check dip, and eventual recalibration.
• The dip is self-inflicted; leaders set expectations too high, then blame the technology when it does not immediately deliver.