Artificial intelligence is reshaping how work gets done, how decisions are made, and how people interact with systems, customers, and each other. The shift from “AI as tool” to “AI as teammate” is already underway. For leaders, however, this shift raises a more difficult question: how do we redesign work in a way that enhances human judgment rather than displacing it?
As a CHRO who has helped organizations navigate major transformations, I see AI not just as a technological challenge, but as an organizational design test. If we’re not deliberate and thoughtful about how we roll out AI in our organizations, the pursuit of efficiency could erode the very qualities (i.e., judgment, trust, creativity) that make human work valuable.
When AI Changes the Nature of Responsibility
According to Microsoft’s research, users of Copilot saved an average of 14 minutes per day. That’s hardly transformational. Yet, too many leaders expect to roll out AI tools and slash headcount as a result. However, companies like Klarna are learning the hard way that jobs cannot simply be viewed as sets of tasks. Rather, most jobs consist of a dynamic interplay of technical skill, human judgment, and emotional intelligence.
On a recent Gartner podcast exploring trends shaping the future of work, a consultant shared the story of a civil engineering firm that tried to implement AI tools into bridge design workflows. The firm’s leadership saw an opportunity for increased speed and accuracy, but the architects resisted. On the surface, it seemed like resistance to innovation. When the consultant dug deeper, however, she discovered that the architects were personally liable, legally and financially, for any structural flaws in the bridges they designed. Introducing a machine into the design process, however sophisticated, posed a moral hazard: if the algorithm made a mistake, the liability still sat with the human.
The consultant wisely recommended drawing a boundary. This wasn’t about digital resistance; it was about understanding where accountability lies. In some situations, it is not only acceptable but necessary to prioritize human judgment over machine assistance.
This story underscores a critical lesson for leaders: Just because AI can be introduced doesn’t mean it should be. Consequently, the real promise of AI lies not in shaving minutes off a task or cutting headcount, but in rearchitecting the very nature of how work is performed.
The Vegas Casino That Got Automation Wrong
Too often, organizations leap into automation without deeply understanding the work they’re trying to optimize. A striking example comes from an unlikely place: a Las Vegas casino.
In an attempt to modernize operations, casino leadership introduced a new feature that allowed slot players to order drinks directly through the slot machine interface. On paper, the idea seemed efficient; it meant cutting down on wait times, streamlining service, and reducing the need for roaming servers. In reality, however, the rollout failed.
Leadership had misunderstood two critical components of the environment:
Customer behavior:
Slot players are famously superstitious. If a machine isn’t paying out, they often get up and move to another. Now imagine someone orders a drink through one machine, switches to another, and a server is left trying to locate them on a sprawling casino floor. What was meant to be efficient became chaotic.
The role of the server:
The job wasn’t just to deliver drinks. Servers played a deeper role, from welcoming guests to reading the room and cutting off patrons who had had too much. These aren’t tasks you can automate; they require human perception, empathy, and judgment. This is what happens when leaders make workflow decisions from a distance, relying on assumptions and dashboards rather than firsthand observation.
In Japanese management philosophy, there’s a concept called Gemba, meaning “the real place.” It refers to the actual site where value is created; examples include the factory floor, the customer interaction, or the design studio. Leaders are encouraged to “go to Gemba” to observe work directly, ask questions, and truly understand the processes and people behind performance.
The casino story is a classic example of what happens when leaders don’t go to Gemba. They optimized for efficiency and lost sight of the human systems that made the customer experience work.
Bots Over Bosses?
Another data point from the same Gartner podcast stuck with me: in a survey comparing human versus algorithmic performance feedback, only 13% of employees disagreed with the statement, “An algorithm would provide fairer feedback than my manager.” Thirty-five percent agreed outright, and the remainder were neutral.
At first glance, this might seem like a glowing endorsement of AI, but I read it differently. This isn’t enthusiasm for automation; it’s disillusionment with the status quo.
And the data supports that interpretation. Only 2% of Fortune 500 CHROs believe their performance management systems work as intended, and just over 20% of employees view performance reviews as fair or transparent. This isn’t a cry for more AI. It’s a plea for more clarity, consistency, and credibility in how human performance is evaluated.
Five Imperatives for Executives in the Age of AI
As AI continues to reshape work, senior executives must ensure its deployment is ethical, strategic, and human-centered.
- Go to Gemba Before You Automate
Before introducing AI into workflows, take the time to understand how work is actually done. Spend time with frontline teams, watch how value is created, and talk to the people doing the work.
This will not only surface hidden complexities; it will help you identify which parts of a process are truly automatable, which require human nuance, and it will prevent you from building technology around false assumptions.
Action: Institutionalize Gemba walks for senior leaders, not as a one-off activity, but as a regular part of strategic planning.
- Redesign, Don’t Just Digitize
Introducing AI isn’t about grafting a new tool onto an old process. It’s about rethinking roles, responsibilities, and workflows. Leaders must ask:
- Who owns a decision when a human and an AI co-produce it?
- Where does oversight begin and end?
- Which tasks need to remain in human hands, and why?
Action: Create joint task forces between HR, IT, and business units to audit workflows and redesign them for augmented, not just automated, performance. The best organizations and leaders treat AI as an interdisciplinary endeavor, not as a technology that is developed and deployed by technologists.
- Fix Performance Management at the Source
If your managers aren’t delivering high-quality feedback, automating the process won’t solve the problem. Before turning to algorithms, address the fundamentals:
- Clear expectations
- Frequent, structured conversations
- Bias training grounded in real behaviors, not check-the-box eLearning
AI can support these systems by surfacing data patterns or summarizing feedback, but it should not become the arbiter of human potential.
Action: Move from once-a-year reviews to ongoing performance and development conversations. Use AI tools as feedback synthesizers, not as decision-makers.
- Build Moral and Data Literacy into Leadership Models
Today’s leaders must know more than how to inspire and execute. They must understand how AI systems work, how data is interpreted, and when to question the output. They must also be fluent in ethical reasoning, particularly in high-stakes contexts where automation could introduce unintended consequences. Increasingly, they must also understand how to tap into the collective wisdom of human-AI teams while preserving the dignity and humanness of their people.
Action: Update leadership competency frameworks to include AI literacy, ethical discernment, and human-AI collaboration as core skills.
- Listen Deeper, Not Just Faster
Don’t rely solely on pulse surveys or engagement scores; ask open-ended questions, and host live feedback forums. Develop AI-friendly cultures by creating dedicated channels for employees to raise concerns about AI tools and how they affect their work.
The best organizations in the AI era will be those that treat employees not as data points but as designers, testers, and stewards of the new world of work.
Action: Launch an “AI + Employee Experience” listening tour to understand how AI adoption is shaping day-to-day realities. Debrief with your leadership team to discuss and debate how you might leverage AI to enhance the employee experience and build, not erode, employer-employee trust.
Designing With Intent
The Vegas casino knew exactly what it wanted from its patrons but failed to understand what its employees were doing to create that experience. In trying to automate, it lost sight of the why behind the work. Now more than ever, leaders must double down on purpose as a source of inspiration, compassion, and profit.
As AI becomes more embedded in our workflows, this is the risk we all face: mistaking visibility for understanding, mistaking speed for effectiveness, and mistaking cost-cutting for progress.
The job of today’s leader is to ensure we don’t make those mistakes, we go to Gemba, we protect human judgment where it matters, and we build organizations where technology enhances the uniquely human aspects of work.
The future of work won’t be shaped by the companies with the most AI. It will be shaped by the companies with the clearest judgment and the deepest understanding of how work truly gets done.
