Introduction
Most workers already believe artificial intelligence is making them more effective on the job. Yet a striking share of employers are not seeing those gains show up in financial results. This gap between individual productivity and organizational profitability is not caused by a lack of data or better algorithms. It is rooted in human behavior, legacy decision-making, and the reluctance to change long-standing rules.
What Happened
Recent findings indicate that 80% of employees report higher productivity when using AI tools. However, only 37% of organizations report any measurable impact on earnings before interest and taxes. In other words, for roughly two out of every three companies, the value employees feel is not yet being captured on the balance sheet.
The issue is not that businesses lack information to make smarter choices. The issue is that knowing what should be done and actually doing it are two different challenges. One case involved a global airline that built a decision-support tool to weigh crew schedules, aircraft rotation, passengers, and ground operations simultaneously. Although the model's recommendations proved valuable, it took eighteen months for controllers to move from accepting about half of them to nearly all. The delay was not technical. Controllers, responsible for passenger safety, had to reconcile each suggestion with decades of professional judgment. The airline ultimately became one of the most profitable in its sector, but expanding the system stalled because of internal hierarchies and unease about an algorithm overriding human experience.
Why This Matters
When employees use AI to complete tasks faster, they are often expected to do even more with the time saved. But if AI is cited as a reason for reducing headcount, demonstrating efficiency can feel like writing one's own redundancy notice. Companies are trimming tasks from existing roles without paying anyone to imagine what those roles could become.
A similar pattern appeared at a global drinks company that combined sales, sentiment, media spend, and weather data by market. When Ireland's rugby team won, spending shifted toward television and Guinness. When the Italian summer lasted longer, a gin brand's season extended. The math worked, but every dollar reallocated changed which country manager would hit a bonus target. The tool ended up reflecting what could have been, rather than driving what actually happened.
This creates what the author calls the incumbent's dilemma at the individual level: protect the past or embrace what comes next. A story about an executive assistant who quietly relied on AI shows that her real value was not in completing tasks, but in reading people, easing tensions, and aligning strong personalities around a shared goal. Technology did not stand in her way; uncertainty about her own worth did.
Multiply that dynamic across an entire workforce, and you get the divide between the 80% and the 37%. The economy's next constraint is no longer something a company can simply raise, build, or buy. This time, the constraint is people.
Key Takeaways
- Eighty percent of employees say AI improves their productivity, but only 37% of organizations see a corresponding EBIT impact.
- The primary barrier is not technical; it is human. Legacy rules, decision hierarchies, and fear of displacement slow adoption.
- Incentive structures matter. When bonuses and promotions still reward old behaviors, employees have little reason to fully trust or act on AI recommendations.
- Leaders should start from an ambitious future state rather than optimizing today, then audit the constraints—including the ones they carry in their own minds.
- Compensation and promotion signals must align with shared results, because employees watch who gets rewarded to understand what truly matters.
- Executives must use the technology themselves. If leaders treat AI as optional while demanding transformation from others, staff will conceal their gains instead of sharing them.
Conclusion
Bridging the AI profit gap requires more than better software. It requires leaders to start from an ideal ambition rather than optimizing the present, to audit the constraints—including the ones they carry in their own minds—and to steer the entire system, not just the budget. Compensation and promotion signals tell employees what truly counts far more than any memo ever could.
Change begins when leadership acts first. Momentum does not start with a plan; it starts with the first honest move. The question for every leader is simple: what will you do differently this week?

Discussion
Join the conversation
Thoughtful reactions, questions, and follow-up ideas help shape the next story.