Andrew Bryant argues that AI-driven efficiency is outpacing leadership capability, creating an “efficiency paradox” where organisations perform better on paper but grow strategically weaker. He explores Klarna’s AI lesson, the shift from performance management to potential development, and why L&D must build leaders who unleash human judgement, creativity, and meaning.
Organisations are getting faster. But they are not getting better at leading people. That is the uncomfortable truth at the centre of the AI revolution. As companies race to automate, optimise, and accelerate, they are discovering that productivity gains come easily. The harder part, the part most are failing at, is developing the human capabilities that make those gains meaningful.
Organisations that chase AI-driven productivity without investing in human development end up operationally successful but strategically vulnerable
I call this the efficiency paradox: organisations that chase AI-driven productivity without investing in human development end up operationally successful but strategically vulnerable (Bryant, 2026). They can process more, analyse faster, and output at scale. But they cannot inspire, adapt, or innovate when the situation demands it. And every situation worth leading through demands it.
The gap nobody planned for
When Klarna deployed an AI assistant that handled 2.3 million customer service conversations, replacing the work of 700 human agents, the metrics were extraordinary. But within months, the company was quietly rehiring humans. The AI was efficient, but not effective. Customers in distress did not want processing, they wanted presence.
Klarna is not an outlier but a preview. Across industries, organisations are discovering that AI excels at replacing human performance but cannot replicate human potential. Performance is what you can measure on a spreadsheet. Potential is what emerges when someone is inspired to go beyond what they thought they were capable of.
The leadership gap opens precisely here. Most leadership development was designed for a world where the primary challenge was managing productivity, getting more output from the same inputs. But AI handles that now. What AI cannot do is create the conditions where people grow, take ownership, think creatively, and find meaning in their work. That requires a fundamentally different kind of leader.
From performance management to potential development
In my research for my book POTENTIAL-IZE I studied hundreds of leaders who have successfully navigated this transition. The ones who thrived had made a critical shift: from managing productivity to unleashing potential. They understood that in the age of AI, the leader’s primary job is not to optimise systems but to develop people.
Tennis coach Tim Gallwey discovered this principle decades ago when he noticed that students improved faster when he stopped giving constant technical corrections. His formula was elegant: performance equals potential minus interference (Gallwey, 1974). The primary obstacle to realising potential is not a lack of ability. It is the mental interference that well-meaning management creates.
When organisations implement AI… managers double down on surveillance and control
This insight has never been more relevant. When organisations implement AI, they often create new forms of interference. Managers, anxious about their own relevance, double down on surveillance and control. Employees, uncertain about their value, become passive and disengaged. The productivity numbers go up, but the human energy goes down.
What is needed instead is what I call the Potentialize Zone, a Goldilocks environment where people are challenged enough to grow but supported enough not to break. In the context of human and AI collaboration, this means calibrating carefully.
- Too cold: letting AI handle everything while humans become passive observers
- Too hot: forcing humans to compete directly with AI on speed and processing
- Just right: using AI to handle routine tasks while humans focus on judgement, creativity, and complex problem-solving
Why L&D must lead this shift
McKinsey’s research on future-of-work skills (Dondi et al., 2021) identified 56 competency elements that will become increasingly important. Self-leadership, interpersonal skills, and cognitive flexibility account for three-quarters of them. Yet most training budgets are still disproportionately weighted towards technical upskilling and tool adoption.
The L&D function has an unprecedented opportunity to close this gap by shifting from a model of skills transfer to one of capability development. The distinction matters. Skills transfer teaches people what to do. Capability development teaches people how to think, adapt, and lead themselves.
The leaders who studied for POTENTIAL-IZE follow, often unconsciously, six interconnected principles I call the IGNITE framework: Inspire, Guide, Nurture, Integrate, Transform, and Evaluate. What makes IGNITE different from traditional competency models is that it treats leadership as a dynamic system rather than a fixed set of skills.
Consider how this applies to L&D design:
- Inspire element is not about motivational posters. It is about helping people author their own story of professional purpose
- Guide is not about prescriptive instruction. It is about building mentoring cultures where the right questions matter more than the right answers
- Nurture means creating psychological safety where people feel they belong and are believed in
- Integrate means teaching people to collaborate with AI rather than compete against it
- Transform means reframing adversity and failure as chapters in a development journey rather than endpoints
- Evaluate means measuring growth, not just performance
Steve Cadigan, LinkedIn’s first Chief HR Officer, predicts we will soon see CVs showcasing learning agility rather than job titles. “We’re going to see more people hired on what they can learn than on what they know,” he told me. “I need to hire for learning velocity.” If Cadigan is right, L&D professionals must rethink not just what they teach but how they identify and measure the capacity to grow.
The real competitive advantage
There is a cautionary tale in all of this. The organisations that treated remote work as a strategy rather than a set of conditions, painfully discovered that aspiration without execution fails every time. The same is true for AI adoption. You cannot optimise for potential by changing only the tools. You must develop the leaders who create the conditions for human beings to do their best thinking, build meaningful relationships, and grow beyond what they thought possible.
Companies have a narrow window, approximately two years, to master the balance between AI capability and human development before competitive disadvantages become structural (Bryant, 2026). Those who wait may find themselves operationally efficient but strategically irrelevant.
The productivity gains are real. But productivity without human leadership creates organisations that can do everything faster and nothing that matters. The gap is not in our technology. It is our ability to lead the humans who use it.
Andrew Bryant is the founder of Self Leadership International and author of POTENTIAL-IZE: How Leaders Unlock Human Potential in the Age of AI and more
References
Bryant, A. (2026). POTENTIAL-IZE: Unlock Potential, Maximize Performance, Inspire Excellence. Wiley.
Dondi, M., Klier, J., Panier, F. & Schubert, J. (2021). Defining the skills citizens will need in the future world of work. McKinsey Global Institute.
Gallwey, W.T. (1974). The Inner Game of Tennis. Random House.
