The Human Leadership Premium Organizations are Actively Devaluing
AI Impact: Organizations are leveraging AI to automate processes but often overlook the essential human capabilities that remain vital.
Leadership Value: Soft skills and human judgment are becoming increasingly important as automation commoditizes other aspects of roles.
Middle Management: The trend towards flattening management structures can erode crucial relational functions that foster team trust.
Long-Term Effects: Removing management layers risks damaging organizational culture and essential interpersonal connections over time.
Strategic HR Role: HR must prioritize preserving relational capabilities rather than solely focusing on cost-cutting efficiency strategies.
There's a strange contradiction running through corporate strategy decks right now. On one slide, the case for AI rests on speed, scale, and the removal of human friction from processes that no longer need it.
On the next, the same leadership team frets about engagement scores, retention of high performers, and a manager population that seems increasingly hollowed out. Both slides describe the same decision.
As organizations push transactional work onto systems that do it faster and cheaper, the capabilities that can't be automated should be appreciating in value. Judgment under ambiguity. The ability to sit with a team through a hard quarter without papering over it. Reading what a customer or a direct report is actually telling you versus what they're saying.
Keep Reading—and Keep Leading Smarter
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Have an account? Log In
These were always part of good leadership, but now they are becoming the part that determines whether an organization functions at all, because the rest is getting commoditized by the month.
Heather Krueger is the Chief People Officer at Engine. She interviews every director-level-and-up hire and has watched the screening criteria move accordingly.
A specialized skill set, certification, or even years of experience doing something a certain way is essentially a depreciating asset. If you’re hiring for what someone knows today, you’re already optimizing for yesterday.
What she screens for instead is the part that doesn't depreciate, such as how a candidate learns, how they handle conflict, what they did when the environment got hard rather than when the press was good. The résumé gets a candidate in the room. The durable human read is what the room is now for. As Krueger puts it:
Learning velocity matters more now than credentials.
The Contradiction
At the precise moment these capabilities are becoming the scarce, expensive thing, a large share of organizations are tearing them out of their own structures.
The post-2023 efficiency wave gave executives permission to flatten, widen spans of control, and thin the middle-management ranks that AI was supposed to make redundant. The logic was that if software absorbs the coordination and reporting work managers used to do, you need fewer managers.
The coordination and reporting work was never the valuable part of the job. The valuable part was the human aspect, the weekly contact that catches a problem before it metastasizes, the coaching that turns a mediocre hire into a strong one, the trust that makes people stay through a rough patch.
When you cut a layer to save on the administrative function, you lose the relational function that was riding on top of it. You don't get to keep half.
The cost shows up on a delay. First the headcount comes out and the quarter looks better. Then the remaining managers, now carrying twelve or fifteen direct reports instead of six, stop being able to do the human part because there aren't enough hours.
The coaching stops. The early-warning system goes dark. Attrition among the people you most wanted to keep ticks up two or three quarters later, by which point no one connects it to the reorg that caused it.
The Case for Letting It Collapse
The strongest version of the opposing view says the middle is supposed to vanish. Kyle Holm, a compensation advisor at Sequoia Consulting Group for 25 years, watches this at AI-native companies.
At AI-native companies, it’s senior people and junior people. There’s no mid-level.
And Holm thinks the rest of the of the economy is headed in the same direction.
His example is sharp. A CEO no longer wants to sit through layers of management to get an idea. Holm's own boss, he says, would rather hear directly from Holm's daughter, a college junior who knows how to channel the tools, than receive a filtered version three levels up.
The young people who harness the technology rise into positions that used to take years to reach. The middle, in this telling, was always just a relay, and the relay isn't needed anymore.
He's right about the relay, but where I tend to disagree with Holm is about what else the middle was doing. Treating the layer as pure idea-transmission is the accounting error that gets companies in trouble.
The reporting function was visible, so it got measured, and when AI absorbed it the layer looked redundant. The coaching, the early read on who's struggling, the trust that takes years to build, none of that showed up, so none of it got counted in the decision to remove it. Holm is describing what disappears. He's under describing what goes with it.
Why this doesn't reverse on demand
The reason this is a strategic error and not just a bad quarter is that the capability doesn't come back when you decide you want it again. Hiring back a layer is easy. Hiring back the trust, the institutional knowledge of who's struggling and why, the relationships that took years to build, is not.
You can post the requisition tomorrow. You cannot post for the eighteen months of accumulated context that walked out the door with the manager you let go. Organizations treat leadership capability as a variable cost they can dial up and down with the cycle. It behaves like a capital asset that depreciates fast and rebuilds slowly.
What the machine can't be accountable for
Grant that AI absorbs the administrative scaffolding around management. The data gathering, the compliance work, the status reporting, all of it. What's left is the part that was always the point.
Ravin Jesuthasan, who has spent years mapping which capabilities survive automation and which don't, draws the line at accountability. There's a category of decision where organizations will insist a person stands behind the call, for reasons of liability, regulation, and ethics, but also for something simpler.
A decision might be 90% done by the machine, but a leader is counted on for her judgment, the experience and expertise to stand behind it and own the outcome. The machine can do the work. It cannot be the thing a board, a regulator, or a frightened employee holds responsible.
The same goes for presence. Jesuthasan expects the work with a genuine human-presence requirement to prove more durable than most of what knowledge workers do, precisely because presence isn't a deliverable a model can hand back. A system can flag that someone is disengaging. It cannot be the person that someone trusts enough to tell the truth to.
The Job of HR and Operations
This is where HR and the operating chiefs have a different job than the one they're often handed. The standard ask is to execute the efficiency plan, to find the headcount, to manage the optics of the cuts.
The more important work is to protect the capabilities that the plan, left to its own logic, will strip out by accident.
That means treating span of control as a quality decision and not only a cost one. It means defending the management layers that do real relational work and being ruthless about the ones that only push paper, which requires actually knowing the difference.
Jesuthasan frames this as a shift HR has been waiting on for years.
You're taking up the mantle of being a steward of work rather than a steward of employment.
The compliance-and-control reflex that defined the function gets handed to the agents. What's left is the architecture of how people and machines actually combine, which is the work that determines whether the efficiency gains hold.
The Automation Funds the Premium
Some organizations are doing this deliberately. They're using AI exactly as intended, to clear the transactional load off managers so those managers can spend the recovered hours on the human work that justifies their existence.
The automation and the human premium aren't in tension in those companies. The automation is what funds the premium. The difference between them and the ones hollowing themselves out isn't access to technology. Both have the same tools. It's ultimately a difference of philosophy and understanding what the tools are for.
The people who run companies are about to spend the next decade learning, expensively, which capabilities they could afford to automate and which ones they were subsidizing through layers they never costed correctly. The bill for getting it wrong arrives late, in attrition reports and stalled execution that traces back to a reorg no one wants to revisit.
Douglas Noll, a lawyer and mediator who trains leaders soft skills, calls the avoidance of this work the empathy paradox. Leaders want the outcomes that empathy produces, performance and retention, but back away from the thing itself because they suspect it makes them look weak. He's blunt about the stakes.
The cost of leaders who can’t connect with their people is low productivity, toxicity, low retention, quiet quitting, and the disengagement companies are already paying for and mostly misdiagnosing.
The capability that's appreciating is the one a generation of managers was trained to treat as soft.
