Following my article on (When Your Manager is an AI agent)
Let me be clear: I’m not anti-AI in management. I’ve spent two years integrating AI into talent development, succession planning, and workforce analytics. I’ve seen the benefits firsthand, and managed myself out of a job by over automating!
But I’ve also seen what we risk losing.
1. Context and Nuance
AI sees patterns. Humans understand context. That distinction matters enormously when you’re making decisions that affect people’s careers, livelihood, and sense of purpose.
2. Development Through Relationship
The best managers develop people through relationship – by knowing them, challenging them at the right moments, believing in them when they doubt themselves. Can an AI do that? Maybe someday. But not yet. And maybe not ever in the way humans need.
3. Meaning-Purpose
Managers help people find meaning in their work, bringing greater investment and discretionary effort to the tasks at hand. They connect individual contribution to larger purpose. They help someone see how their struggle today builds the capability they’ll need tomorrow. AI can tell you what to do. But it can’t tell you why it matters to you specifically, based on who you are and who you’re becoming.
4. Moral Judgment
Management requires moral judgment, not just performance optimization. When do you give someone grace? When do you hold firm? When does someone deserve another chance despite the data? When does the data reveal a truth the person can’t see themselves? These aren’t algorithmic decisions. They’re human ones.
5. The Unquantifiable
Some of the most valuable contributions can’t be measured:
- The person who lifts team morale
- The colleague who makes others braver
- The mentor who helps someone believe in themselves
- The person who asks the question that changes how everyone thinks
If management becomes purely about optimizing measurable performance, we systematically devalue everything that can’t be measured.
And we end up optimizing ourselves into in-humanity.
The Real Question
So here’s what we need to ask:
Not “Can AI manage people?”
But “Should AI manage people?”
And more specifically:
“What parts of management should remain irreducibly human, no matter how good AI becomes?”
I think the answer looks something like this:
AI Should Handle:
- Performance tracking and reporting
- Resource allocation optimization
- Skills gap identification
- Workload balancing
- Compliance monitoring
- Pattern recognition across large datasets
- Administrative task management
Humans Must Own:
- Building trust and psychological safety
- Contextual judgment calls
- Meaning-making and purpose connection
- Moral and ethical decisions
- Development through relationship
- Navigating ambiguity and complexity
- Holding space for human vulnerability
- Inspiring people through difficult change
The question is: Can we maintain that boundary? Or will we drift – slowly, imperceptibly – toward algorithmic management because it’s more efficient, more scalable, more “objective”?
What Happens to Managers?
If AI handles the “manageable” parts of management, what becomes of managers themselves?
Scenario A: Elevation
Managers are freed from administrative burden to focus on the irreducibly human parts of leadership. They become coaches, culture-builders, meaning-makers, strategic thinkers. The AI handles the performance tracking. The human handles the performance growth. This is the optimistic future. And it’s possible.
But it requires something crucial:
Organizations must value and reward the human capabilities of management, not just the measurable outputs. If we still promote people who hit numbers but erode culture, we’ll get managers who treat AI recommendations as gospel and ignore the human elements.
Scenario B: Elimination
Middle management largely disappears. AI handles the operational management. Senior leaders manage AI managers. Individual contributors are either managed by AI or are senior enough to be self-managing. The layer in between – the human managers who used to translate strategy into execution and support people through the work – becomes redundant. We’ve been predicting the death of middle management for decades. AI might actually deliver it.
Scenario C: Subordination
This is the darkest timeline. Managers remain, but they’re effectively managed by the AI systems they’re supposed to be managing. Their decisions are constrained by algorithmic recommendations. Their judgment is overridden when it conflicts with data. Their performance is measured by how well they implement AI suggestions, not how well they develop people. They become human interfaces for algorithmic management.
The title says “manager” but the role is “AI Implementation Officer”
The Manager-AI Power Dynamic
Here’s something nobody’s talking about:
Who manages the managers when both are AI?
If you’re a human leader managing AI managers who manage humans, you have a span of control problem. The AI managers generate massive amounts of data. You can’t possibly review it all. So you rely on summary reports. Aggregated metrics. Exception flags.
Basically, you rely on AI to tell you what the AI managers are doing.
See the problem? You’re not managing AI managers. You’re being fed information by AI systems and making decisions based on what they choose to show you.
This is algorithmic control disguised as human leadership.
A Personal Confession
I need to be honest with you. I’ve been part of creating this future. I’ve built AI-enhanced succession planning systems. I’ve implemented learning analytics that track every intervention. I’ve designed performance management processes that generate data-rich dashboards for leaders.
And I’ve seen the benefits:
- Better identification of high-potential talent
- Earlier intervention for struggling employees
- More equitable promotion decisions
- Reduced bias in performance reviews
These are real gains. But I’ve also seen the drift. I’ve watched leaders stop talking to their teams and start reading dashboards instead. I’ve seen managers defer to AI recommendations even when their gut says something’s wrong. I’ve observed organizations optimize for what’s measurable and lose sight of what matters.
And I wonder: Am I building the tools that make managers more effective, or am I building the tools that make managers obsolete?
The Choice Ahead
Here’s what I believe: We’re at a fork in the road.
Down one path: We let AI do more and more of management because it’s efficient, scalable, and removes human bias and inconsistency. We end up with algorithmic management systems that optimize performance but lose the human elements that make work meaningful.
Down the other path: We deliberately choose what must remain human in management, and we fiercely protect those boundaries. We use AI to handle the administrative and analytical parts of management, freeing humans to focus on development, meaning-making, and relationship.
The difference between these paths isn’t the technology.
It’s the choices we make about what we value.
If we value efficiency above all else, we’ll get algorithmic management. If we value human development, meaning, and dignity, we’ll use AI as a tool while keeping management fundamentally human.
What Leaders Must Do Now
If you’re leading organizations through this transition, here’s what you need to do:
1. Draw Clear Boundaries
- Decide explicitly: What parts of management will always be human?Write it down.
- Make it policy.
- Don’t let it drift.
2. Redesign Management Roles
If AI handles performance tracking, what do your managers actually do? Redefine the role around the irreducibly human capabilities, and resource it appropriately – you can’t expect managers to do deep coaching work while also carrying 47 direct reports.
3. Measure What Matters
If you only measure and reward quantifiable performance, you’ll get managers who optimize for that. Find ways to value and reward the unquantifiable human contributions.
4. Develop Human Capabilities
Invest in developing managers’ coaching, empathy, judgment, and meaning-making capabilities. These are the skills that matter when AI handles the analytics.
5. Create Transparency
If AI is making or influencing management decisions, make it visible. People deserve to know when they’re being managed by an algorithm.
6. Maintain Human Override
Always allow human judgment to override AI recommendations. And create a culture where doing so is acceptable, not seen as resistance to data.
7. Ask the Ethical Questions
Just because AI can manage people doesn’t mean it should. Keep asking: What are we losing? What are we risking? Where’s the line?
The Future We Choose
I don’t know whether we’ll end up with human managers leading AI managers, humans managed by AI, or some hybrid we haven’t imagined yet.
But I know this:
The future of management won’t be determined by what AI can do.
It will be determined by what we decide humans should do.
What I Hope For
I hope we build futures where AI makes managers better, not redundant. Where technology handles the tracking so humans can focus on the developing. Where data informs judgment but doesn’t replace it. Where efficiency gains are reinvested in human connection, not used to eliminate human managers. Where we measure performance and we value the unmeasurable contributions that make teams come alive.
I hope we choose to keep human relations more human.
Responses