The Intelligence Shift: Why AI Adoption is a Heart Game, Not a Tech Problem

Brice Challamel, Head of AI Strategy and Adoption at OpenAI, argues that successful AI integration is not a technical challenge, but a fundamental shift in human agency. The core idea is that we are moving from knowledge work, which is the manual synthesis of information, to intelligence work, where AI acts as an operating layer for human judgment. The hidden consequence of this transition is that organizational failure is rarely about the technology itself, but about the heart game: the emotional and psychological readiness of leaders to engage with non-deterministic systems. Readers who grasp this distinction gain a competitive advantage by avoiding the trap of treating AI as a deterministic tool, instead positioning themselves to lead the intelligence shift while others remain paralyzed by the fear of an alien, uncontrollable system.

The Hidden Cost of Deterministic Thinking

Most organizations approach AI as they would traditional software: a tool that follows rigid, predictable rules. Challamel suggests that this mindset creates a structural barrier to adoption. Leaders who excel in deterministic environments, such as those who spend hours fine-tuning Excel macros or managing top-down, command-and-control teams, often struggle when faced with AI.

If you enjoy iterative relationship building with children and cats, which is very tentative at first requires a lot of humility. It is going to fail as often as it succeeds but it is going to take you to a great place. Then you are probably more ready than not to welcome AI in your life.

-- Brice Challamel

This reveals a critical systems dynamic: AI is not a tool to be commanded, but a partner to be negotiated with. When leaders force deterministic expectations onto non-deterministic models, they experience the AI as a juggernaut of alien intelligence. The downstream effect is a failure to build the necessary trust. Trust, Challamel notes, cannot be dictated; it must be earned through thousands of iterations. Organizations that fail to shift from commanding to iterating will find themselves stuck in a loop of frustration, blaming the technology for their own inability to adapt to its probabilistic nature.

The Long First Mile and the Myth of Completion

Challamel introduces the concept of the ten-step journey, where the final 10% of the work, the part that actually delivers impact, is often the most difficult because it requires the highest level of clarity. The hidden risk here is exhaustion-driven failure. After completing nine steps, leaders are often too depleted to recognize that the original plan may no longer be relevant.

The 90% that you have done is most likely the condition to do the last 10%, which is the work that actually matters. But when you get there, you are exhausted because you have already done 90% of the work.

-- Brice Challamel

This creates a competitive trap: teams that are reactive to the initial plan rather than the current reality will fail the final step. The advantage goes to those who treat the tenth step as a separate, high-stakes phase requiring a mental reset. If you are not prepared to walk back your assumptions from the first nine steps, you are likely to execute a plan that is fundamentally obsolete.

Compassion as a Systemic Strategy

Perhaps the most non-obvious insight is that the battle for AI adoption is won through empathy, not argument. Challamel argues that most resistance to AI is not rational; it is a projection of past trauma and existing existential dread. When leaders attempt to convince naysayers through logical demonstration, they are engaging in a process that literally translates to tying with ropes.

The system responds to this coercion with resistance. Because the opposition is rooted in the heart rather than the mind, the only effective intervention is to listen first. By lowering the emotional temperature, leaders can uncover the specific silent struggles that drive fear. This is not just a soft-skills observation; it is a tactical necessity. In an environment where AI is increasingly viewed as a threat to professional identity, the ability to lead with curiosity, rather than debate, is what differentiates successful organizations from those that face constant, cyclical pushback from their own workforce.

Key Action Items

• Audit your Deterministic Bias: Over the next quarter, identify processes where you demand 100% predictability from AI. Shift your focus toward iterative, conversational workflows where you treat the AI as a partner rather than a tool.
• Implement a Reset Mechanism: When reaching 90% completion on a major AI project, force a 48-hour debrief and pivot period. Do not execute the final phase until you have re-verified that your initial assumptions still hold true in the current environment. This prevents the exhaustion-driven errors Challamel warns about.
• Lead with Curiosity, Not Coercion: If you encounter resistance to AI, stop debating. Dedicate time to listen before you think. Use the next 12-18 months to map the specific silent struggles, such as job security, identity, or fear of obsolescence, driving the pushback.
• Focus on the Intelligence Briefing: Move your team away from manual knowledge work, such as reading, scanning, or drafting, toward intelligence work, such as briefing, recommending, or deciding. This requires building internal agentic workflows where agents surface signals, allowing humans to focus on judgment.
• Earn Trust Through Small Demos: Do not attempt to dictate trust in AI. Over the next 6 months, prioritize hundreds of small, low-risk demonstrations of AI capability to build organizational confidence before scaling to high-stakes, sensitive tasks.