AI is taking over the tasks managers used to fill their day with. Their role isn't disappearing — it's being redefined. Most haven't been told what their new job actually is.
Walk into a leadership meeting at almost any organization rolling out AI right now and ask the question: What is the manager's job in 18 months?
You'll get nervous silence. Maybe a few generic phrases about "strategic thinking" and "human connection." But almost no one can articulate it concretely. That's a problem — because the answer determines whether your middle management layer becomes your biggest AI accelerator or your biggest source of resistance.
Harvard Business Review's 2025 analysis on how AI is redefining managerial roles makes the case directly: tasks that once consumed large portions of managers' days can now be done automatically. The role itself isn't being eliminated — it's being reshaped. What managers are responsible for is changing, even when the title on the org chart stays the same.
"Tasks that once consumed large portions of employees' days can now be done automatically — reshaping what managers and employees are responsible for, not eliminating the roles themselves." — Harvard Business Review (2025)
That reframing matters. The question isn't "will my job exist?" It's "will I be doing the work my organization actually needs in 18 months — or will I still be doing the work AI just absorbed?"
Translate information between layers. Gather data, summarize it, send it up. Take direction from above, distribute it down. Run status updates. Coordinate workflows. Schedule, route, and approve.
Translate ambiguity into direction. Develop judgment in their team. Hold people accountable to outcomes, not activity. Coach through uncertainty. Build trust across functions. Make decisions AI cannot.
The shift isn't subtle. The old job was mostly about moving information. The new job is mostly about developing humans. AI handles the first. It cannot do the second.
PwC's data tells leaders something most haven't fully internalized: the managers who develop AI fluency aren't just becoming more productive. They're becoming structurally more valuable to the organization — and to the market. Which means the managers who don't develop it aren't just falling behind. They're becoming replaceable in a way that compounds quickly.
That's the quiet career risk most middle managers haven't been told about. And it's the quiet opportunity that organizations should be telling them about — clearly, and now. (For the financial case behind investing in people first, see The 70% Rule: Why People, Not Technology, Determine AI ROI.)
The people who got promoted into management over the last decade got there largely because they were good at the old job. They were efficient at processing information, reliable at coordinating workflows, and dependable at running the operational machinery.
None of those skills automatically translate to coaching, judgment development, or holding difficult conversations about performance and growth. Some managers will make the transition naturally. Many won't — not because they're not capable, but because no one has told them the job is changing. And as I covered in What Employees Are Actually Afraid Of When It Comes to AI, the silence around what's changing is exactly what fuels resistance.
Redefine the role explicitly. Most job descriptions for managers haven't changed in a decade. They list responsibilities that AI is about to absorb. If the description doesn't change, neither will the expectations — and neither will the behavior.
Train for the new skills, not the old ones. The manager development programs most organizations run are about budgets, project management, and operational metrics. The new programs need to be about coaching, feedback, and developing judgment in others. That's a different curriculum.
Measure managers differently. If you're still evaluating managers on the volume of work they push through, you're measuring the old job. The new metrics are team capability growth, decision quality, and how effectively their people use AI to operate at a higher level.
Give them permission to use AI themselves. A manager who is uncomfortable with AI cannot coach their team to use it well. The first AI users in your organization should be the managers, not just the individual contributors. Most organizations get this backwards.
The organizations that get this transition right will have managers who become force multipliers — coaching teams to operate at a level that wasn't possible before. The organizations that don't will quietly hollow out the middle layer, because managers without a clear new job will either leave on their own or be eliminated when the headcount math gets done.
Neither outcome happens by accident. Both are the result of choices leaders make right now, in the next 12 months, while the tools are being rolled out.
The AI Leadership Workshop helps managers see what's changing — and what to do about it — before the gap opens up. Built for organizations actively deploying AI.
See the workshop →