Entry-level jobs used to double as leadership school. The scheduling, first-draft writing, and data-pulling that junior staff did for years quietly taught them how organizations actually work, long before anyone called it leadership training. As AI absorbs that work, HR leaders are confronting a leadership pipeline that is thinning out just as it becomes harder to refill.
The rungs AI is quietly removing
Nobody scheduled a training session called “how to read a room” or “how to spot which numbers matter.” Junior employees learned those things by doing unglamorous work: drafting the first version of a report a manager would later mark up, sitting through meetings to take notes, pulling and cleaning data before anyone trusted them with an opinion on it. Each task was a low-stakes rep at a skill that later became indispensable in a leadership role: synthesis, judgment, stakeholder awareness.
Generative AI tools now do a growing share of that first-draft, first-pull work. That is precisely the point of adopting them, and few HR leaders would argue for keeping it manual purely for training’s sake. But the tasks were never just busywork. They were the reps. Remove enough of them and the informal curriculum that built middle managers, and eventually senior leaders, starts to disappear without anyone deciding to cancel it.
What the data says
That is the concern driving a new guide from Talogy, a global talent management solutions provider. Talogy surveyed 207 senior HR leaders, talent acquisition managers, and learning and development experts, and found that 78% are concerned about the long-term loss of critical leadership skills as AI takes on tasks and responsibilities traditionally associated with entry-level roles. The guide, drawing on research involving more than 6,000 employees, managers, and talent professionals worldwide, was authored by Ali Shalfrooshan, Talogy’s Director of Science Strategy.
“As organizations rethink employee development, teaching employees how to work alongside AI tools simply isn’t enough,” Shalfrooshan said. “We must be far more intentional about cultivating core human capabilities, such as critical thinking, adaptability and collaboration if we are to enable people to take on complex responsibilities over time. Leadership development needs to begin at the early career stage and evolve dynamically throughout the entire career journey.”
The four capabilities Talogy says matter most
The guide groups the human capabilities it argues organizations must now cultivate deliberately, rather than let develop on the job, into four areas it calls RISE: Responsiveness, the openness and agility to learn and adapt as work evolves; Insight, the ability to think critically and make sense of information; Sustainability, the capacity to stay resilient and sustain performance through change; and Effectiveness, the ability to communicate and collaborate to turn ideas into outcomes. None of the four are technical skills. That is deliberate: the guide’s argument is that as AI absorbs technical and procedural tasks, the human capabilities around judgment and collaboration become the scarcer, harder-to-automate differentiator, and they are also the ones least likely to develop without a manager or a program designed to grow them on purpose.
A two-track hiring market makes it worse
The leadership-pipeline concern does not sit in isolation. Separate labor-market research from PwC’s 2026 Global AI Jobs Barometer, which analyzed more than a billion job postings across 27 countries, found entry-level roles splitting into two tracks: postings in heavily AI-exposed occupations have grown even as traditional entry-level postings in the same fields have shrunk, and the entry-level roles that remain increasingly demand skills, like stakeholder management and independent judgment, that used to appear only later in a career. In other words, the jobs that still exist at the bottom of the ladder are being redefined to require the very capabilities that bottom-of-the-ladder jobs used to be where people first developed them.
That squeeze compounds the problem Talogy’s survey respondents are describing. If entry-level work increasingly expects judgment before it teaches it, and mid-level work increasingly automates the low-stakes reps that used to build that judgment, the traditional assumption that leadership capability accrues naturally with tenure stops holding.
What it means for the HR leader
The fix Talogy’s guide points to, and what shows up across the leadership-development conversation more broadly, is to stop treating leadership capability as something that develops by osmosis and start treating it as something to design for explicitly. That has practical implications for how HR teams should be running L&D and succession planning right now, not eventually. HRTech has covered how heavily some leadership decisions already lean on AI output without a human check built in, and the same instinct, letting a tool quietly substitute for a developmental step rather than support it, is what is emptying out the pipeline here too, as covered in AI Efficiency Is Quietly Costing Managers Their Judgment.
Three moves are worth prioritizing. First, audit which entry-level tasks in your organization have been substantially automated in the past 18 months, and ask explicitly what skill each one used to teach. Second, replace the lost reps on purpose: structured stretch assignments, shadowing, and deliberately assigned first drafts that a manager still reviews, rather than assuming employees will find equivalent learning opportunities on their own. Third, use validated psychometric and behavioral assessment, not just performance history, to identify leadership potential earlier and route development resources to it, since tenure alone is becoming a weaker signal of readiness than it used to be.
None of this works if development stays informal while the informal path disappears. HR teams that are already stretched thin on AI skills-building for their workforce, a gap this publication has covered before in Employers Are Free-Riding on Workers’ AI Skills, now have a second, related bill coming due: the informal leadership curriculum AI is quietly dismantling has to be rebuilt on purpose, or the bench simply keeps shrinking without anyone noticing until it is too thin to promote from.
Source: Talogy
