Companies are under growing pressure to move quickly on artificial intelligence. New tools are being evaluated, vendors are being brought in, and AI initiatives are increasingly appearing on executive agendas.
But what if the biggest obstacle to successful AI transformation isn’t the technology?
For Yusra Mahmood, Founder of EchoData and Chief Business Officer at NAZZTEC, many organisations are confronting problems that existed long before generative AI entered the boardroom: slow approvals, unclear ownership, unreliable data, and leaders who haven’t clearly defined the decisions they are trying to improve.
Working with CEOs, boards, and executive leadership teams on AI and transformation decisions, Mahmood takes a business-first approach focused on commercial impact, operational readiness, and measurable outcomes.
Her central argument is straightforward: most companies don’t have an AI problem. They have a decision problem.
Highlights
What If Your AI Problem Isn’t About AI?
One of the earliest mistakes, Mahmood says, happens when executives treat AI transformation as a procurement exercise.
“Honestly, the biggest mistake is treating AI like a procurement decision,” she says. “Leadership sits in a room, picks a vendor or a tool, announces it to the company, and calls that a strategy.”
The harder question comes before selecting the technology:
What decisions in this business are we bad at making, and why?
Without answering that, companies risk applying AI to organisational problems they haven’t resolved.
“Most companies don’t have an AI problem,” Mahmood says. “They have a decision problem that’s been there for years: slow approvals, no clear ownership, data nobody trusts. AI just makes that dysfunction visible faster.”
She has encountered executives eager to introduce a chatbot before leadership had even agreed on who internally owned customer data.
For Mahmood, that contradiction captures a much larger problem with the rush towards enterprise AI.
“You can’t automate a decision that nobody is willing to own.”
Six Months Later, Leadership Asks: Where’s the ROI?
The problem doesn’t end once a company identifies where it wants to use AI.
Mahmood also sees organisations push responsibility for transformation down to technical teams, only for executives to return months later wondering why those projects haven’t generated meaningful returns.
“The other mistake is skipping the leadership layer entirely and pushing AI down to the technical teams,” she says. “Then six months later, leadership is confused about why there’s no ROI.”
Technology teams can implement systems, but leadership still has to determine what business outcome those systems are expected to produce.
For Mahmood, that means executives themselves have to be willing to change.
“AI transformation has to start at the top,” she says, “with leaders willing to change how they personally make decisions—not just with a tool rollout to IT.”
It is a distinction that shifts AI transformation away from being solely a technology initiative. The question becomes not only what the organisation is implementing, but whether its leaders are prepared to operate differently because of it.
The Leaders Who Know the Most About AI May Not Win
That raises another question: what kind of leader succeeds when AI becomes embedded in more business decisions?
Mahmood doesn’t believe technical expertise will necessarily be the defining advantage.
“I think the leaders who thrive won’t be the ones who know the most about AI technically,” she says. “They’ll be the clearest thinkers who can define a problem precisely, know what a good decision actually looks like for their business, and are comfortable being accountable for a call even when a model helped make it.”
AI creates two very different temptations for leaders.
Some may delegate judgment too readily, treating AI as a black box they don’t need to understand. Others may refuse to change how they operate at all.
“Both are avoidance, just in opposite directions,” Mahmood says.
Her advice for preparing is therefore less about mastering the technology itself and more about examining the quality of one’s own judgment.
“The way to prepare isn’t a technical course,” she says. “It’s getting sharper about your own decision-making—how you weigh evidence, where you’re biased, what you actually optimise for.”
That becomes particularly important when a model contributes to a decision without removing the executive’s responsibility for its consequences.
As Mahmood puts it:
“AI rewards leaders who already know how to think clearly. It exposes leaders who don’t.”
Final Thoughts: When AI Gets Smarter, Who Owns the Decision?
The promise of AI is often speed: faster analysis, faster processes, faster decisions.
But speed becomes far less valuable when an organisation isn’t clear about where it is going.
Mahmood’s perspective points to a less glamorous prerequisite for AI transformation. Before companies ask what technology can automate, leadership has to decide what deserves to be automated, what outcome it should improve, and who remains responsible when a decision is made.
That last question may become more important as AI becomes more capable.
Executives will increasingly make decisions with systems that can analyse, recommend, predict, and automate. Yet greater technological capability doesn’t necessarily make human judgment less important. It can make the quality of that judgment more consequential.
The companies that benefit most from AI, then, may not simply be those that adopt the fastest.
They may be the ones whose leaders know what they’re trying to achieve before the technology enters the room.
Highlights
Read the Chinese article here.
Related posts:
He’s Led AI Transformations Across 10 Countries. His Award-Winning Projects Weren’t the Most Advanced.
AI Workforce Transformation in Asia: The Silent Revolution
AI Is Saving Your Team Time. Your Company May Be Taking It Right Back.
AI Can Now Work Across Departments. Your Company Is Still Organised Around Them.
