The Human Side of AI

Across the Middle East, AI budgets are large and the ambition is real. The UAE has a dedicated AI ministry. The dashboards look healthy. But in organization after organization, the transformation the spending promised has not shown up.

That gap was the subject of the second episode of The Lean Scale, a conversation with Dr. Regis Chasse, Chief Learning Officer at CLO Advisors. His argument is that the constraint is rarely the technology. It is the leadership and the system around it, and whether either is ready to be redesigned.

Chasse has spent two decades on enterprise transformation: co-leading Capgemini University, running the leadership institute at Majid Al Futtaim during its pivot against Amazon's entry into the GCC, and leading the global leadership practice at Heidrick & Struggles. That last role produced a two-year study across 350 organizations. Its finding is counterintuitive: as AI is embedded more deeply, organizations become more human, not less. When AI absorbs transactional work, what remains is the work only people do well, collaboration, judgment, empathy. Those skills stop being a supplement and become the point.

AI adoption is not a transformation

Chasse separates three relationships an organization can have with AI. He credits the framing to Arizona State University.

Understanding AI is awareness: what it can do, what it cannot, where the risks are. It changes nothing on its own.

Using AI is where most people work today, by Chasse's estimate around ninety percent. Prompting, drafting, checking a decision, rehearsing a conversation. It raises individual productivity and leaves the process untouched.

Creating with AI is where almost no one operates. It means building new workflows and deciding, deliberately, what the AI does and what the human must. This is the only level where the word transformation is accurate.

Most investment stops at the second step. Tools get bought, individuals get faster, and that speed gets mistaken for change. The process underneath stays the same.

What changes when you redesign the work

Take a process every organization runs: performance management. The employee writes a self-evaluation, the manager adds feedback, and a calibration meeting settles on a rating.

Using AI improves the edges. The employee drafts a sharper self-evaluation. The manager cleans up the feedback. Both rehearse the hard conversation with an AI coach beforehand. The process is untouched.

Creating with AI changes the process. An agentic system works from objectives set at the start of the year and collects evidence as the year goes: results in dashboards, moments recorded in email, work as it actually happened, rather than a manager reconstructing twelve months from memory in the final week.

That removes the recency effect, where the last month outweighs the other eleven. It replaces memory with continuous evidence. And it changes what the people do: the manager and employee now review the evidence, challenge it, and catch a problem mid-year while it can still be fixed.

The harder part is not the technology. It is the willingness to redesign the system around it, and to become more human inside it rather than fearing replacement by it.

Deloitte's 2026 Human Capital Trends research found that organizations taking a technology-first approach to AI are 1.6 times more likely to miss the returns they expected than those taking a human-centered one.

Four principles for human and AI collaboration

If the case is that AI investment needs a parallel investment in leadership, these four capabilities are what that investment buys. All four are human.

Trust

AI creates fear of being replaced. Trust is what brings people into the change instead of leaving them defending against it. It rests on confidence in four things: the intent behind using AI, the process, the data, and the governance. Without it, people perform adoption without ever using AI well. In the GCC, where business already runs on trust, this is a familiar demand made new: it is built through what a leader does daily, not what they announce.

Judgment

The easy failure is what Chasse calls cognitive laziness: the recommendation looks right, so it passes unchecked. But accuracy is a known limit of these systems. The human job is to judge the output, whether the evidence is relevant, what it means, and what is missing. The same holds for compensation, where AI now supplies the benchmarks: the number cannot set the reward on its own.

It is no longer the era of knowing. It is the era of learning. Satya Nadella, cited by Chasse

Accountability

Accountability is the refusal to hide behind AI or process. The system can supply every input to a decision, but a person still owns it, on quality and on consequences. In a performance review, the manager cannot hide behind the collected data and the generated text. The question stays human: is it fair?

Learning agility

AI changes weekly; a capability that appears one morning can be gone the next. No redesign is right the first time, so feedback loops have to be built into it. A team might find that how a manager phrases objectives in January decides what the system captures all year, or that collaboration is going unmeasured. Learning agility is not training people on the tool. It is examining how well the new process works and improving it.

The new operating system for leadership

Naming what matters more is half the shift. The other half is naming what matters less.

The first thing to give up is being the expert. Many leaders were promoted for what they knew, not how they led. Expertise is now available to everyone in the organization, so being the most knowledgeable person in the room is a fading source of authority.

The second is control. An organization that insists on controlling every flow of information cannot run agentic AI. The shift is from policing information to improving outcomes. A leader who is less the expert, less in control, and more human-centered is the one who lets trust, judgment, and accountability take hold.

Change the environment, not just the leader

So how do these four become the way an organization actually works? Not through a three-day course. Everyone has seen a leader return from an excellent program, energized, and change nothing.

Chasse explains why. The leader learns empowerment, comes back motivated, and tries to apply it, but the environment was never built for it. The systems do not support it. Their own manager holds the old view. The mindset moved; the environment did not. The change dies.

Leadership development cannot sit apart from the transformation. It has to be built into it.

In a real transformation the work runs in parallel streams, people, process, technology, data. Leadership belongs inside the people stream, at the table, often through the chief people officer, governing alongside the rest. The system changes, and leadership changes as part of it.

What this means for learning leaders

For CLOs, the measure of success is no longer how many leaders a program reaches. It is whether the mindset shifts land in step with the rest of the transformation, matched to the use cases being built. The conversation moves from the cost of training to the capabilities a given use case needs and the value it creates.

Human-led, powered by AI

• The why. AI spending needs a matching investment in leadership. Focus only on the technology and the returns underdeliver.
• The what. Four capabilities carry the next few years: trust, judgment, accountability, learning agility. Hire and develop for these. They are no longer soft skills.
• The how. Build a system, not a program. Tie the capability to the strategy and embed the learning in the work.

Behavior changes when mindset changes, and both have to connect back to the strategy. The future is human-led and powered by AI. The work of leadership is to focus on the human part.