Maybe It’s Not the Engineers Who Should Worry About AI

It’s time we ask the uncomfortable question: What if it’s not developers who should be worried about AI replacing them… but managers?

I had two conversations recently with senior leaders in two completely different industries—healthcare and education. Neither knew the other. Different org structures. Different goals. Same takeaway:

Middle management is broken.

And not just flawed or underperforming—doomed to be ineffective by design. In both cases, these leaders observed that middle managers were often only validated by the very anti-patterns that harm company efficiency and team health. It’s not that they were bad people—it’s that the role, as structured, rewards the wrong behaviors.

🧱 The Middle Management Trap

Let’s be honest. Many middle managers were promoted because they were great individual contributors—not because they knew how to lead. They get sandwiched between high-level directives and low-level execution, with very little room to think, question, or innovate.

Instead of clearing paths for their teams, they’re:

• Translating vague direction into over-planned roadmaps,
• Measuring output over outcomes,
• Prioritizing appearances over progress,
• Scheduling check-ins and surfacing “status,” but not removing blockers.

They’re caught in a performance loop that has little to do with real work and everything to do with looking busy.

🤖 What If AI Replaces Managers, Not Engineers?

Everyone keeps talking about AI replacing developers.

But if we’re being real: AI is already doing what many middle managers do—but better.

• Aggregating status updates ✅
• Tracking progress ✅
• Surfacing team health signals ✅
• Recommending process changes ✅
• Sending reminders and running retros ✅

AI doesn’t play politics. It doesn’t protect turf. It doesn’t waste your time with a meeting that should’ve been a Slack message. And it doesn’t promote people for managing optics.

Suddenly, it’s not so hard to imagine a world where AI augments or even replaces layers of management—especially the ones who exist to validate a system, not question or improve it.

🌀 Bureaucracy Isn’t Just Inefficient—It’s Dangerous

This hits especially hard in fields like healthcare and education, where bloated org structures are literally suffocating progress. We’ve created layers upon layers of approval, alignment, and oversight, but fewer people with real context, accountability, or agency.

Every layer between the people doing the work and the people making decisions adds noise. That noise turns into inefficiency. That inefficiency gets justified as complexity. And we call it management.

But it’s not leadership.

💡 The Managers Worth Keeping

This isn’t an anti-management rant. Good managers do exist. They:

• Develop their people
• Coach instead of control
• Build clarity instead of process
• Connect strategy to action
• Advocate up, not micromanage down

They lead by removing themselves as bottlenecks, not inserting themselves into every decision.

But here’s the hard truth: most orgs reward the opposite.

🪞 AI Didn’t Create the Redundancy—It Just Made It Obvious

AI didn’t invent bureaucracy. It just made it easier to see through the illusion. If a bot can replace your “value add,” maybe it wasn’t that valuable.

And ironically, the engineers building those bots might be the safest—not because they can’t be replaced, but because they’re close to the problem. They’re solving something. Making something. Building something real.

Management isn’t going away. But the layers that only exist to preserve power, validate hierarchy, or perform productivity? Those might not survive.

And maybe that’s a good thing.