The International Monetary Fund warned in 2024 that artificial intelligence could disrupt nearly 40% of jobs worldwide. Advanced economies face even steeper exposure as AI transforms both routine tasks and specialized work.
Boardrooms across the globe have responded swiftly. The strategy is familiar: adopt AI faster, boost efficiency, cut labour costs, and rebuild operations around automation.
The pressures driving this shift are genuine and measurable. AI is already reshaping software engineering, customer service, financial services, logistics, legal review, cybersecurity, and strategic decision-making at scale.
Companies that stand idle risk losing competitive ground. Yet a critical problem is emerging—one most organizations haven't seriously addressed.
Many firms are redesigning their operations quicker than they understand resilience itself. Board intervention here must take a different, harder approach.
Most AI strategies assume human labour is the main cost to eliminate. Real-world experience tells a different story entirely.
Microsoft scaled back parts of its AI coding initiative when costs spiralled beyond projections. Uber burned through its entire 2026 AI coding budget in just four months.
What seemed like simple labour replacement proved far more complex. The technical, financial, and architectural challenges caught organizations off guard.
AI economics work nothing like traditional software economics, it turns out. Deploying advanced AI agents means token consumption skyrockets rapidly.
Infrastructure costs climb steeply as well. Monitoring, validation, and human review demands rise sharply with each new system.
Research shows some AI tasks consume up to 1,000 times more tokens than standard interactions. Costs fluctuate wildly even for nearly identical work.
In many cases, running AI systems costs more than paying humans for the same jobs. That doesn't kill the AI opportunity—it rewrites it completely.
Winners won't be companies that automate most aggressively. They'll be organizations that redesign work with genuine intelligence, blending AI power with human judgment and institutional knowledge.
Yet board-level conversations on this remain dangerously immature. Governance discussions haven't caught up with the reality on the ground.
Most boards still measure AI through cost cuts and productivity gains alone. Important metrics, yes—but they're incomplete pictures.
Real strategy demands a harder question: Is your organization building or destroying the foundations of long-term resilience? Human intelligence remains part of that foundation—perhaps the part most underestimated.
Chasing pure efficiency risks discarding uniquely human capabilities. Interpreting ambiguity, handling edge cases, making ethical choices—these matter deeply.
Understanding context, building trust, responding to unprecedented events—AI can't reliably do these things. Resilience gets built or lost exactly here.
Organizations that will outperform aren't those removing humans most ruthlessly. They're the ones architecting human-AI work deliberately and strategically.
