The rise of artificial intelligence and shifting workplace dynamics like flatter organizational structures have sparked intense debate about the future role of managers. Eight industry experts weigh in on whether middle management is an endangered species or an evolving necessity in modern organizations. Discover their insights on which leadership capabilities matter most when humans and machines work side by side.
Architect Decisions Across Intelligent Systems
AI isn’t making managers obsolete. It’s making one specific type of manager obsolete: the one whose entire value was controlling information flow.
For decades, middle management existed because information was scarce and hierarchy was the routing system. Managers moved data upward, translated strategy downward, and approved decisions in between. AI collapses that function quickly, and it should.
But what replaces it isn’t a simpler structure. It’s a more demanding one.
Organizations that are genuinely AI-native don’t just have fewer managers. They need managers who can design and govern the intelligence layer itself. Who owns the escalation thresholds? Who decides which decisions stay with the system and which require human judgment? Who ensures accountability doesn’t disappear when a workflow becomes automated? That’s not a lighter job. It’s a harder one.
At Automations24, we call this “decision architecture.” In one engagement with an operations leadership team, removing manual approval layers didn’t simplify their management structure; it raised the skill floor. They suddenly needed people who could interpret what the system was doing, intervene at the right threshold, and redesign the workflow when operating conditions changed. The managers who remained weren’t the ones who knew the process. They were the ones who understood the logic behind it.
The organizations I watch struggle aren’t the ones that eliminated management layers. They’re the ones that automated execution without building governance around it. That’s not a technology failure; it’s a management design failure.
AI doesn’t make management less important. It makes bad management design immediately visible.
Paul Malott, CEO, Automations24, Inc.
Elevate Trust Through Human-Centered Leadership
Ineffective managers are becoming obsolete in the age of AI and flatter orgs, but management is more important than ever.
We can all see AI optimizing workflows, automating processes, and synthesizing data. What it can’t do is sit as a human, across the organization, and support employees and customers, specifically when it comes to building trust, psychological safety, and managing change.
Trust is the operating system of any high-performing team, and it is built in human moments, not dashboards.
Early in my HR career, I worked with a team whose manager was technically brilliant but relationally absent. Productivity metrics looked fine on paper. What the data didn’t show was that three of his top performers were preparing to leave, and one was in crisis. At this point, there is no algorithm that can replace effective conversations and listening with the people actually doing the work.
The moment we replaced him with a manager who prioritized psychological safety over output pressure, people started talking. Problems surfaced early. Retention stabilized. The work improved.
Flatter structures and AI tools don’t eliminate the need for someone who can hold the human center of a team. They raise the bar for what that person has to be. The managers who survive and thrive in this era won’t be the ones who were good at pushing through tasks. They’ll be the ones who were good at making people feel seen, heard, and safe enough to do hard work.
That is not a soft skill. That is the skill.
Stephanie Lemek, Founder & CEO, The Wounded Workforce
Own Accountability For High-Stakes Judgment
Managers are not becoming obsolete. What is disappearing is the part of the job that was never really management to begin with.
Flat structures and AI tools mostly attack one function: moving information up and down a hierarchy. That was real labor, and machines do it faster, but it was never the actual job. The actual job shows up in the moment a system produces a confident answer that happens to be wrong, or the situation is ambiguous and no procedure covers it. Someone has to own the call of what to do and how to do it. Accountability is the one thing you cannot flatten or hand to a model. You can spread a decision across 10 dashboards and one algorithm, but when it goes wrong, a regulator, a customer, or a court will still be looking for the single human who owned it.
The reason is volume. AI does not reduce the number of consequential decisions an organization makes. It multiplies them and speeds them up. The more decisions are made faster, and with more machine influence, the more moments that need a person who can be held responsible. That makes the right kind of manager more valuable than before. The manager whose entire role was passing information along is the one in trouble.
My example comes from aviation. Over 50 years, cockpits shed crew and added automation. The flight engineer position vanished. Autopilots took over most of the flying. By the obsolescence logic, the captain should have been next. Instead, the captain’s authority became concentrated. We keep a properly knowledgeable and competent human in command precisely for the rare, unscripted moment the automation cannot handle, and we hold that person formally accountable for the outcome, no matter how much the machine did along the way. Every aviation safety management system still names an accountable executive who cannot delegate that responsibility away. The same principle is now arriving inside every organization that puts AI in the path of a real decision.
Philip Mann, Principal Consultant, Vector Strategic Consulting LLC
Fuel Motivation With Close Relationships
The work managers do is more important than ever. Early in my career, I hit the manager jackpot and was lucky to learn from someone who cared about my success. The work was stressful, with long hours and a never-ending list of customers to engage. And yet, my manager was deeply human. She created space where I could evolve professionally and trusted me to make and execute detailed plans for projects that had never been done before. She asked thoughtful questions, gave sound advice, and believed I would grow into a capable leader.
Intrinsic motivation is the fuel behind human productivity. A key driver of that motivation is relationships. When we feel a sense of connection and belonging, we have more energy to do great work. While I find immense value in AI as a tool to support business success and even my own personal productivity, its value is limited by my ability to remember to use it and craft strategic prompts.
The harsh reality is that humans are complex creatures, each with unique personalities, interests, and life goals. When we expect groups of individuals to get work done together, there will always be different perspectives—and those perspectives come with vastly different emotional experiences and coping strategies. For many logic-driven business leaders, this may be a tough pill to swallow, but the fact is, humans need other humans to thrive.
No matter how flat organizations become or how capable AI tools grow to be, managers are not going anywhere. After all, the trope that people leave bosses, not businesses, has a lot of truth behind it.
Kate Vawter, Author, Better Boss Blueprint
Restore Order Via Disciplined Hierarchy
Managers are more important than ever, and their importance will only grow in the coming years. I’ve been responsible for scaling orgs and leading transformational change. Hierarchy shouldn’t be a dirty word—it’s the lack of basic communication that makes it problematic. Flatter org structures don’t democratize orgs, they magnify chaos and dysfunction. And they generally emerge only when an org is rife with lousy managers.
With hierarchy, you can have structured one-on-one meetings with your manager that address areas to improve, opportunities for growth, and other things that require continual monitoring. AI has become remarkably good at efficiency gains around repeatable tasks. But I’m unconvinced AI can manage the discernment and delicacy required to talk to people about their performance or development. We might get there soon, but we’re not there now.
When I joined the Consumer Financial Protection Bureau as a founding chief of staff, the tech team was the CIO and myself. We scaled it to over 250, half of them remote across the country, long before remote work was the norm. That growth wasn’t possible without hierarchy. Flat structures sound inclusive—everyone has a voice, everyone weighs in—but in practice, they’re deeply inefficient at scale. Decisions stall. Accountability blurs. The same conversation gets repeated in five different rooms because no one knows who owns the call. What made a distributed team of 250 function wasn’t flatness; it was managers who knew their people, understood their work, and could make judgment calls in real time. AI couldn’t have built that. Neither could a flat org chart.
We have a serious management problem—we are awash in mediocre management. We continually promote talented individual contributors to management positions, when the two skills are entirely different. Management requires a deep understanding and appreciation for people. Being a great data scientist doesn’t make you a great manager, and yet we promote fantastic contributors with rarely any training—let alone good training.
Rachael Goldfarb, Cofounder, The Coul & Gold Group
Translate Mission Into Daily Purpose
AI can make us more efficient, and flatter structures reduce bureaucracy. But managers become even more critical, and their roles shift to being meaning-makers. Alongside AI, they manage purpose, not tasks.
When you flatten a company, AI tracks the projects and does the quality assurance, but you still need to create an environment where talented people stay because they want to be connected to the company’s mission, not just because they need the paychecks. So the manager then takes on the role of translating the high-level purpose of the company into the low-level everyday purpose for their team. We’re seeing employee experience becoming one of the top metrics for talent acquisition because candidates are evaluating culture and leadership before they enter.
The best managers are able to create such an environment by creating internal storytelling that connects the dots and contextualizes the work being done above and below key performance indicators.
And within flatter consumer brand companies I know, managers use meetings not for status updates (which are done by AI) but to discuss stories about how particular ground-level interventions have caused community initiatives to be funded more. By constantly connecting execution to the larger goal of the company, they significantly changed the culture and reduced turnover rate.
Encouraging intrapreneurial spirit and connecting what people do every day to the broader impact of the company is a uniquely human managerial skill that cannot be taken by software.
Zach Dannett, Cofounder & Co‑CEO, Tumble
Amplify Superstars Amid MultiThreaded Stewardship
Single-domain/single-threaded managers will be in lower demand as AI reduces coordination costs and increases access to baseline cross-functional expertise. The people who can become “supercontributors” not bounded by any particular discipline grow even more effective; they simply don’t need that direct management layer providing regular oversight, or task management, or low-level alignment.
Multithreaded managers who handle multiple streams of efforts simultaneously are just as valuable as ever, primarily because they allow the supercontributors to maximize their own efforts and outcomes.
Even in my role, I’m directly managing five independent streams of work that have only one or two senior people assigned to them, with no team-level manager. The senior contributors responsibly empowered with AI dramatically extend their breadth and depth of execution, and the management job becomes mitigating conflict, political support, or ensuring higher-level decision alignment rather than managing day-to-day project tasks or morale.
Flatter orgs with AI can be turned into a huge impact advantage for the managers who can adapt.
Joseph Gefroh, SVP of Product and Engineering, HealthSherpa
Enable Collaboration Beyond Positional Authority
AI is making one kind of manager obsolete: the ones whose authority stems from hierarchy. The question worth asking is why that model became so dominant in first place—and so persistent.
For decades, many managers profited from information asymmetry. Their higher rank in the org charts gave them access to management meetings where data, context, decisions, and reasoning about why they were taken were exchanged—insights that others did not have. This asymmetry justified that it is they who decide. This hierarchy makes sense for industrial business models where companies increase profits by making workers more efficient. More control creates value.
Now, AI is revolutionizing how every employee can access information, which undermines the information asymmetry. The remaining justification for a manager’s role is what they enable in others.
Modern digital solutions need input from various domains—more than any single manager can cover. The key to success is collaboration and exchange. Teams flourish in environments where members feel safe to speak up. Managers who rely on hierarchies to justify their ambition to decide and control, in contrast, are undermining this competitive advantage we have over machines.
What makes positional authority so persistent in most organizations goes deeper than their industrial roots. For most of human history, communities relied on natural leveling mechanisms. When leaders stopped benefiting them, the group had ways to push back. Modern hierarchies removed these leveling mechanisms. Leadership is now selected top-down and accountable to who stands above in the org chart. The one-pager culture most office workers know firsthand is a symptom of exactly that. The relevant knowledge sits with the people doing the work, and they are best-placed to make decisions—but the system routes decisions upward anyway.
Command-and-control managers who insist on having the last word are becoming a liability—it’s those managers who create the conditions for the right word to surface who will survive.
Dr. Dominik Hörndlein, AI Strategy & Implementation Consultant | Author, Dominik Hörndlein Consulting
