Something has shifted in the last few months.
In our conversations with tech leaders and CEOs of tech-enabled businesses, the tone is noticeably different. AI is no longer ‘emerging’ – it’s here. And with it comes a burning question: what kind of leader do you need to be when AI runs through the fabric of your organisation?
According to McKinsey’s latest State of AI research, around 88% of organisations now report using AI in at least one business function. Yet only around a third have begun to scale it in a way that delivers material enterprise-wide impact.
Meanwhile, Morgan Stanley research on 935 firms across five AI‑exposed sectors finds that companies with at least one year of AI adoption report an average 11.5% increase in net productivity over the past 12 months, alongside a 4% net reduction in headcount.
The gains are real. But so are the implications for leaders, their people, and the way organisations are structured.
The World Economic Forum’s Future of Jobs Report 2025 estimated that 86% of employers expect AI and information processing technologies to significantly transform their business by 2030.
Those companies on the front foot are not just adding AI tools to existing structures – the most ambitious are rethinking their operating models entirely.
One CEO we work with describes a quiet but substantial commitment: a 20% headcount reduction by year-end, “across all roles”, with the explicit mandate to “rebuild our company as if we were founding it today”. The method: identify 15 to 20 distinct roles in the organisation, clarify what each role is there to do, and then ask honestly what proportion of that work can now be redefined or automated.
But two problems surface repeatedly in our conversations with CEOs and boards.
The first is cognitive overload of people. AI expands what’s possible – generating ideas, presenting options, accelerating thinking. But it also introduces a new kind of pressure. People across organisations – including leaders themselves – need to be skilled and confident in deciding what matters and where their judgement should be prioritised.
Boston Consulting Group research supports this: while over three-quarters of leaders and managers now use generative AI several times a week, regular use among frontline employees has stalled at around 51% – partly because the cognitive demands of integrating AI into workflows are proving harder than expected.
The second problem is time. AI is, at this stage, genuinely time-consuming to use well. People are spending time exploring tools, learning what works. Without guardrails, it can eat up the working day and pose more problems than it solves.
A different kind of leadership is needed to navigate this new world, and it’s becoming clear that, in this new context, those leaders who are thriving are not simply the most technologically literate.
One Alexander Partnership client describes them as “leaders who combine experience of building and leading in the analogue world with genuine comfort in the AI one”.
They are rare precisely because those two qualities have rarely needed to coexist before.
Conventional leadership qualities still matter. What’s changed is that there is a new layer on top: the ability to lead leaders who are themselves leading AI, against a backdrop of ambiguity.
What does that mean in practice? It means thinking in systems. It means framing the right problems. It means understanding where human judgement is irreplaceable and where it is now redundant.
Some leaders we work with are experimenting deliberately in practical ways to embed AI into everyday work. One operates a “one-thirds rule”, expecting employees to spend at least a third of their time using AI to solve problems.
Others run short AI sprints, where small teams are given a defined project or process redesign to deliver using only AI tools, focusing less on the outcome and more on understanding what works.
But what happens to the established, experienced CEO whose comfort with change has been outpaced by the speed of AI – and who has not yet invested the time to understand what it means for how they lead?
We hear of some funds and investors who are beginning to issue explicit mandates: leaders who are not engaging with AI seriously are at risk of being replaced.
Our concern is that replacing seasoned leaders too quickly, in favour of more tech-native successors, risks losing hard-won leadership capability alongside the outdated habits.
Experience still matters. Judgement still matters. The question is whether those qualities are accompanied by sufficient adaptability.
We see the leaders who are navigating this well sharing something in common. They are not waiting for clarity before acting. They are comfortable with the ambiguity, letting go of what has gone before and embracing new ways of doing things.
They are learning in public, making space for experimentation, and treating their own development in AI as seriously as they treat their organisation’s.
They are modelling the behaviour they expect from everyone else.
In a recent conversation, one CEO we work with summed it up:
“I think boards are looking for CEOs who can handle disruption and ambiguity better than their predecessors. Boards also like to have the certainty of someone who has experience… but the reality is that often they really need someone who has the agility and humility to adapt to the new world.”
So, what does this mean in practice?
Right now, these are the most honest questions a leader can ask themselves:
- Am I genuinely fluent in AI, or am I observing it from a polite distance?
- Have I made a deliberate, structural decision about how AI changes the way my organisation works – or am I hoping it will evolve naturally?
- Do I understand the difference between deploying AI tools and redesigning the work itself?
- Am I creating the conditions for my people to develop real fluency – with protected time, clear mandates, and space to experiment and fail?
And perhaps most importantly: am I willing to go first?