Stop Building AI Labs. Start Solving Business Problems

“So… where exactly is the AI button?”

A CIO asked me that, half-jokingly, during a strategy session. Everyone laughed. But the confusion behind the question was real.

In boardrooms across the world, I’ve seen the same pattern emerge. Executives know they need to “do AI.” Budgets are approved. Labs are built. Pilots are launched. But real transformation? Still out of reach for a lot of organisations.

Despite the tools and talent, something isn’t clicking. Billions have been spent, yet few initiatives have scaled beyond controlled environments or made it to the press releases. According to MIT Sloan, 70% of AI pilots fail to reach production. That number should concern all of us, not because AI lacks potential, but because we’ve misunderstood where the real value begins.

The Real Problem Isn’t AI, It’s How We Approach Work

The uncomfortable truth is this: we’ve been treating AI like a software deployment instead of what it is, a business reinvention. Too often, it’s disconnected from the very processes and people it’s meant to empower. We fall in love with the technology and forget to ask: What is the work we’re trying to transform?

A colleague recently shared a habit that stuck with me. He said, “Every day, with everything I do, I ask myself, can AI help me do this faster, better, or smarter?” That simple question reflects a mindset shift I believe we urgently need. It’s not about chasing the next big platform. It’s about examining the work itself and asking where friction lives, and how we can redesign it.

From Intelligence to Impact, Agentic AI in Action

This is where a different kind of AI thinking enters, Agentic AI.

Agentic AI doesn’t try to be human. It doesn’t write poetry or mimic creativity. It performs clearly defined tasks, within boundaries, at scale. It’s embedded into daily workflows, not abstracted into innovation centers. Whether it’s a finance agent reconciling thousands of transactions or a compliance assistant drafting regulation updates, Agentic AI works inside the system, quietly and reliably.

The beauty of this approach lies in its simplicity. No manifestos, no moonshots. Just AI that does real work, accelerating outcomes, reducing manual load, and freeing people to focus where human judgment matters.

Dubai Didn’t Wait, They Executed One Task at a Time

In October 2017, the UAE became the first country in the world to appoint a Minister of State for Artificial Intelligence. It wasn’t a symbolic gesture, it was a statement of intent. AI wasn’t going to be treated as a tech experiment. It was to become a national engine of transformation.

Let me acknowledge my bias: I was born and raised in Dubai. And I’ve had the privilege of watching the city evolve, from sand to skyline, from standard infrastructure to 3D printed offices. Few places move from vision to execution with the same clarity and speed. Dubai didn’t wait to perfect a roadmap. They acted, then refined. While many organisations were still drafting strategies, Dubai was already embedding AI into immigration, healthcare, traffic optimisation, and public licensing. These weren’t loud platform rollouts. They were measured, outcome-driven, and grounded in execution. Agentic by nature, business-led by design.

That’s what makes the agentic mindset so powerful. It doesn’t require a five-year roadmap. It requires the courage to act, and the discipline to start small, but smart. Today, I see many enterprises appointing roles like Head of AI, GenAI Leads, or Automation Program Directors, which is a great sign. But it also raises an important question worth reflecting on:

Where should this role sit, and what is it accountable for?

Should it live inside IT or Innovation or with Strategy? Or should it have a broader, more integrated mandate, one that connects business and delivers the new paradigm!

There’s no single answer. But here’s what I’ve learned:

When AI is treated as a function, it stays narrow. When it’s viewed as a business enabler, it scales.

It is essential to remember that AI ultimately unlocks business value and enables enterprise-wide change. That only happens when ownership is shared, impact is measured, and value is aligned across the organisation.

Why So Many AI Programs Stall

What I’ve seen, and what data continues to reinforce, is that many AI programs fail because they chase ambition without ownership. They start with tech, not with a business problem. They lack KPIs, they lack shared responsibility between business and delivery, and they bolt on governance as an afterthought. In some cases, the foundational trust in automation is missing altogether.

Which brings me to a lesson that shaped my thinking more than a decade ago.

🧠 Lee’s Rule: If Humans Don’t Need to Touch It, They Shouldn’t

Back in 2009, I was working at a bank in the Middle East when our CIO, Lee North, introduced a principle I’ve never forgotten:

“If humans don’t need to touch it, they shouldn’t.”

He was referring to production changes, automating them end-to-end with scripts, exception handling, and audit trails.

That wasn’t AI. But it was the mindset that made true automation possible, and scalable, inside that bank. It was about trusting systems, designing with intent, and building a culture that didn’t glorify manual heroics, but instead valued consistency, clarity, and scale.

And here's the funny part: while it took multiple tools duct-taped together, a lot of after-hours effort, and some coffee-fueled debugging marathons... The very program I thought might never see the light of day ended up becoming a case study, and went on to win awards.

To this day, I see organisations eager to “do AI,” but still hesitant to automate the basics.

Final Thought: Literacy Over Labs

We don’t need just more AI centers of excellence. We need a workforce that understands where AI belongs, and where it doesn’t.

We need boards that know how to ask the right questions. We need teams that can spot the opportunity inside the work they do every day.

You don’t need a five-year AI roadmap. You need five well-chosen tasks, reimagined through an agentic lens, this quarter.

Let’s stop chasing artificial intelligence. Let’s start building agentic intelligence, one business problem at a time.

And yes — in case you’re wondering — AI helped me shape this article. It didn’t write it for me. But it did help me clarify, structure, and refine my thinking. I won’t pretend otherwise, because that’s exactly the point:

The real value of AI isn’t in replacing our voice, it’s in helping us express it better, with more purpose and precision.

#AITransformation #AgenticAI #FutureOfWork #DigitalStrategy #EnterpriseAI #Leadership #BusinessTransformation #AIExecution #ANZLeadership