Guest Blog by Jonathan Bowder & Joe Folkman
AI is a leadership problem disguised as a technology problem.
What makes leaders matter when AI can do so much of the work? As AI absorbs analytical and executional tasks, the defining question shifts from “what can AI do?” to “what can only leaders do?” The CEOs of JPMorgan, Microsoft, Citadel, and Klarna have each said some version of this publicly in the past year. Everyone is asserting it, but no one has tested it against a large-scale leadership assessment database – until now.
We turned to our database of more than 2 million assessments of 138,777 leaders to identify the leadership skills that remain uniquely human – the ones AI has not, at least not yet, been able to replicate. To keep the data relevant, we limited our analysis to leaders evaluated from 2023 onward, capturing the period in which generative AI moved from novelty to workplace fixture.
1. The coaching paradox. As AI absorbs the doing, coaching becomes the manager’s primary job – the very capability most leaders were never promoted for. When we asked leaders to rank the importance of Developing Others against 19 other leadership competencies, they placed it 11th. Direct reports ranked it 10th. Not unimportant, but not treated as essential either.
The data tells a different story about its actual impact. Leaders who score in the bottom quartile on Developing Others are rated, on average, in only the 21st percentile for overall leadership effectiveness. Leaders in the top quartile land in the 85th percentile. That’s one of the widest effectiveness gaps in our competency model, tied to a skill leaders consistently underrate.
AI can support an employee’s development – surfacing resources, drafting a learning plan, even simulating a difficult conversation. What it can’t do is notice that someone is ready for a stretch assignment, deliver hard feedback with enough trust behind it to land, and follow up weeks later to reinforce or correct course. That ongoing loop of observation, feedback, and follow-through is still a manager’s job, and it’s becoming the job, not a side task fit in around “real work.”
2. The motivation tax. Across the more than 15,000 leaders we’ve assessed since 2023, Inspires and Motivates Others was the lowest-rated of our 19 competencies. Yet when direct reports ranked which competencies matter most to their success, Inspires and Motivates came out on top. By contrast, Drives for Results ranked 4th in both effectiveness and importance.
Put together, this is a clear picture of leaders who are good at push and weak at pull – effective at driving for results, far less effective at inspiring the people who have to deliver them. AI is very good at the push: it can model targets, track progress, and optimise a plan. It cannot supply the meaning, energy, or belief that gets a team to want to hit that plan. As execution gets automated, the scarce leadership currency shifts from driving work to igniting it.
3. Adoption runs on trust. Trust in the manager isn’t a soft metric – it’s a leading indicator of retention and effort. When trust is low, 45% of direct reports are thinking about quitting, only 22% are willing to put forth extra discretionary effort, and just 27% would recommend the organisation as a good place to work. When trust is high, those numbers flip: only 18% think about quitting, 57% offer extra effort, and 62% would recommend the organisation.
Our research points to three sources of trust:
→ Positive relationships – built through genuine interactions and shared interest, not transactions.
→ Consistency – between what a leader says and what a leader does.
→ Expertise – confidence that comes from a leader’s demonstrated knowledge and experience.
People are learning to trust AI for what it knows. But trust in a leader is built differently – through relationship and consistency accumulated over time, not just accurate output. AI can have the facts. It doesn’t have the track record.
This isn’t an isolated finding. In a related analysis of over 100,000 leaders, we compared people who scored in the top quartile on Champions Change, Innovates, and Technical/Professional Acumen against their peers to see what else separated top AI-transformation leaders from the rest. Four more competencies stood out: Develops Strategic Perspective, Establishes Stretch Goals, Communicates Powerfully and Prolifically, and – again – Inspires and Motivates Others. The overlap is the point: the skills that make a leader effective at driving AI adoption are largely the same skills that make a leader effective, period. AI didn’t invent a new leadership model. It raised the cost of neglecting the human one that was already there.
That reading is backed by how employees themselves feel about the shift. Recent workforce research from Workday found that 83% of people believe AI will make human skills more important, not less, and 76% say they want deeper human connection as AI becomes a bigger part of their work lives. Employees aren’t bracing for leaders to become more technical. They’re bracing for leaders to become more human.
None of these three findings – coaching, motivation, trust – are new ideas in leadership research. What’s new is the leverage. For decades, developing others, inspiring a team, and building trust were “nice to have” alongside the technical and operational work that consumed most of a leader’s time. AI is now taking a real bite out of that operational work. What’s left in the leader’s job description is disproportionately the human part – and the data shows most leaders have under-invested in exactly that part for years.
That’s the real risk in the AI transition. It isn’t that leaders will be replaced by a model. It’s that the skills separating average leaders from extraordinary ones – coaching, inspiring, earning trust – will matter more than ever, right as many organisations pour their development budget into technical AI fluency and assume the human skills will take care of themselves. They won’t. Our data shows a 64-percentile swing in leadership effectiveness tied to one underrated competency alone.
The leaders who thrive in this next decade won’t be the ones who out-compete AI at analysis or execution – that’s a race they can’t win and shouldn’t try to. They’ll be the ones who double down on what was always the harder, more human half of the job: growing people, motivating them, and earning their trust one consistent action at a time. AI will keep getting better at the doing. The leading has never been more essential – or more up for grabs.
Joe is Co-Founder of Zenger Folkman. A globally renowned psychometrician, heholds a doctorate degree in Social and Organizational Psychology, and a master's degree in Organizational Behaviour from Brigham Young University. He is a best-selling author or co-author of nine books on leadership and feedback.
Jonathan is a certified trainer of NLP, a Certified Master Coach, a Certified Zenger Folkman Extraordinary leader facilitator as well as qualified in organisational change from Geert Hofstede’s ITIM organisation. He is also one of the UK’s few Zenger Folkman Extraordinary Leader Master Trainers.
