Organizations are spending $366 billion a year on leadership development—and most of it isn’t working. As AI reshapes every function, the gap between the leaders companies have and the leaders they need has never been wider.
The modern leadership pipeline is breaking under compound pressure. Seventy-seven percent of chief human resources officers lack confidence in their bench strength for critical roles, and only 20% have successors ready to step into key positions, according to the DDI Global Leadership Forecast 2025—the largest study of its kind, spanning 10,796 leaders across more than 50 countries. At the same time, 88% of organizations have embedded AI into at least one business function, per McKinsey’s 2025 State of AI survey, and the World Economic Forum projects that 39% of core workforce skills will change by 2030. This is not a slow-moving talent gap. It is a structural mismatch between the leaders organizations are developing and the leaders the AI era demands—and closing it has become the defining challenge for every C-suite executive and HR leader today.
The pipeline is hollowing out from the inside
For three consecutive years, Gartner’s annual survey of more than 1,400 HR leaders has ranked leader and manager development as the number-one HR priority. Yet the investment is not translating into results. Gartner’s own research found that traditional leadership development—seminars, lectures, and episodic workshops—actually produces a negative effect on development outcomes. Seventy-five percent of organizations rate their programs as “not very effective,” even as more than half have increased spending.
The problem runs deeper than pedagogy. DDI’s 2025 data reveals a workforce where 71% of leaders report increased stress and 40% of those stressed leaders have considered leaving their roles entirely to protect their wellbeing. Trust in immediate managers has plummeted to just 29%, a staggering 37-point decline since 2022. Meanwhile, high-potential individual contributors—the very people organizations are counting on to become tomorrow’s leaders—are heading for the exits. Their intent to depart rose from 13% in 2020 to 21% in 2024, and those whose managers fail to provide regular growth opportunities are 3.7 times more likely to leave within the next year.
Deloitte’s 2025 Global Human Capital Trends report, drawing on nearly 10,000 business and HR leaders across 93 countries, adds another dimension to the crisis. Managers currently spend nearly 40% of their time on problem-solving and administrative tasks, leaving just 13% for developing their people. The irony is sharp: at the exact moment organizations need managers who can coach, inspire, and guide teams through AI-driven transformation, those managers are drowning in operational busywork.
Finding tomorrow’s leaders in a sea of data
For decades, identifying high-potential leaders has relied on frameworks like the 9-box grid—used by an estimated 67% of Fortune 500 companies—which plots employees on a matrix of current performance against future potential. Korn Ferry’s Assessment of Leadership Potential evaluates seven signposts including learning agility, drivers, and derailment risks. DDI has defined ten high-potential factors ranging from authenticity to conceptual thinking. These frameworks have produced real insight: Korn Ferry’s research demonstrates that individuals with high learning agility are promoted twice as fast as their peers, and organizations with highly agile executives enjoy 25% higher profit margins.
But the traditional model has serious blind spots. Eighty-six percent of HR leaders acknowledge bias in their succession planning processes. Manager nominations—still the primary gateway to high-potential programs in most companies—are heavily influenced by proximity, personal relationships, and unconscious preference for candidates who mirror existing leadership archetypes. Career interruptions for caregiving or health unfairly suppress ratings. And annual talent reviews provide only static snapshots of a dynamic workforce.
AI is beginning to change this calculus. The share of organizations using AI for HR functions jumped from 26% to 43% between 2024 and 2025, according to SHRM, and platforms like Eightfold AI, Pymetrics (now part of Harver), and SHL are enabling continuous, data-rich talent assessment at scale. Pymetrics’ neuroscience-based gamified assessments measure more than 90 cognitive and emotional traits with completion rates of 98% and accuracy above 89% in skill matching. Internal talent marketplaces from companies like Gloat algorithmically match employees to stretch assignments that reveal potential in real time, replacing the once-a-year calibration session.
The promise is significant: organizations using AI-powered recruitment report 31% faster hiring times and a 50% improvement in quality of hire. Properly implemented AI reduces hiring bias by 56 to 61% across gender, racial, and educational categories. Yet the risks are equally real. Amazon’s widely documented abandonment of an AI recruiting tool that systematically discriminated against women remains a cautionary tale. New York City’s Automated Employment Decision Tool law, enforced since July 2023, now requires independent bias audits before AI can be used in hiring. The emerging consensus—reflected in McKinsey’s latest thinking—is a hybrid evaluative model: AI handles pattern recognition and scale while human judgment retains authority over values, cultural context, and creative potential.
Why the old development playbook fails in an AI-augmented world
If identification is being transformed by technology, so too is what comes after it. The global leadership development market exceeds $366 billion annually, with roughly $166 billion spent in the United States alone. Organizations that invest well see returns: research estimates suggest $7 in value for every $1 invested in leadership development, with companies reporting 25% better business outcomes and 2.3 times higher cash flow per employee. Yet nearly 60% of first-time managers receive no training when they transition into leadership, and 75% of learning and development professionals estimate that less than half of what they teach actually gets applied on the job.
The AI era is forcing a fundamental redesign of how development happens. AI-powered coaching platforms are democratizing access to what was once a C-suite privilege. Coachello, which integrates AI coaching into Slack and Microsoft Teams, delivered a 37% reduction in time to leadership readiness at Carrefour and an 83% increase in manager confidence in eight weeks at other client organizations. Valence’s AI coach Nadia, customizable to each company’s culture and values, produced a measurable 5% uplift in leadership effectiveness scores. McKinsey itself has deployed a generative AI platform called Lilli to 30,000 users and embedded “AI black belts” on teams, with entry-level associates now using AI tools from their first day.
Virtual reality is scaling beyond pilot programs. A PwC study found VR-trained employees completed training three times faster than classroom learners and were up to four times more focused than e-learners. Accenture’s persistent virtual campus reduced caseworker turnover by 31% and cut training costs by 75%. A 2025 Forrester and Meta study documented 219% ROI from enterprise VR training with payback in under six months. Gartner now estimates that by 2025, 75% of employees in new roles will be trained or coached by AI first—not by a person.
The shift is from event-based to continuous development. Rather than pulling leaders out for a week-long program, organizations are embedding micro-coaching, AI-driven simulations, and peer learning directly into the flow of daily work. BCG’s behavioral science-rooted methodology integrates leadership development into daily routines. The evidence increasingly supports what Gartner has made explicit: traditional seminars are not just ineffective—they may actively impede growth by creating a false sense of development.
The competencies that define leadership now
What, then, should these new development systems actually teach? In a landmark January 2026 article, McKinsey global managing partner Bob Sternfels, World Economic Forum president Børge Brende, and senior partner Daniel Pacthod identified three uniquely human leadership domains that AI cannot replicate: setting aspiration and enrolling others to own it, demonstrating judgment aligned to values when those values conflict, and designing for nonlinear outcomes by fostering creative dissent and breakthrough thinking. Their argument is direct: “AI may transform how we work, but only human leaders can determine why we work and what we’re trying to achieve.”
Harvard Business Review reinforced this framework from a practical angle. In an October 2025 article, London Business School professors Herminia Ibarra and Michael Jacobides outlined five critical skills leaders need: cultivating AI fluency through diverse networks, redesigning organizational structures to unlock AI’s value, orchestrating human-AI collaboration, coaching teams through AI-driven change, and leading by example in adoption. Their research noted a troubling gap: while 374 S&P 500 companies mentioned AI on earnings calls—nearly nine in ten in glowing terms—most defaulted to vague references to “productivity gains” when pressed for specifics. Only one in three leaders genuinely understands key AI concepts, according to Harvard Business Impact’s 2025 study of more than 1,100 learning and development professionals.
The data converges on a paradox at the heart of modern leadership. McKinsey projects organizations will need 26% more social and emotional skills by 2030, even as analytical roles shrink. Ninety-nine percent of organizations in the Harvard Business Impact study identify emotional intelligence as essential for leadership, and nearly half say its importance has grown in just the past year. TalentSmart’s research shows emotional intelligence explains 58% of job performance and that 90% of top performers score high on EQ. Yet Korn Ferry finds that only 22% of leaders actually demonstrate real emotional intelligence today.
The leaders who will thrive are not those who choose between technological fluency and human depth. They are those who fuse both. As IBM’s research projects, 74% of executives expect AI to redefine leadership roles enterprise-wide by 2030, with two-thirds anticipating entirely new AI-driven leadership positions. IMD professor Mark Greeven captures the shift: “The winners will not be those deploying the most models, but those reinventing how decisions, teams, and accountability are organized around AI.”
The window for action is narrowing
The confluence of a hollowed-out leadership pipeline, a $366 billion development industry producing mediocre results, and the fastest technological transformation in a generation creates an urgent imperative. Senior leaders and CHROs cannot afford incremental adjustments. Four actions matter most right now.
First, redefine what potential looks like for an AI-augmented future—shifting from credentials and past performance to learning agility, adaptability, and the capacity for human-AI collaboration, using validated assessments augmented by AI analytics with rigorous bias audits. Second, replace event-based development with continuous, embedded learning that meets leaders in their daily workflow through AI coaching, VR simulation, and peer networks. Third, invest in emotional intelligence and trust-building with the same rigor applied to technology budgets—these are not soft skills but the hard differentiators that determine whether transformation succeeds or stalls. Fourth, protect leadership capacity by using AI to eliminate the administrative burden that currently consumes 40% of managers’ time, freeing them to do what only humans can: inspire, coach, and make judgment calls when values collide.
The WEF’s message from Davos 2026 offers the right frame: “The future will belong not to those who go the fastest, but to those who bring the most people with them.” In the age of AI, leadership is not being replaced. It is being revealed as the one capability no algorithm can substitute—and the one most organizations have chronically underbuilt.
