When I started out thinking about AQ and the impact of managing people and agents in my last article (How AQ will define the next generation of leaders) I realised we’ve been asking the wrong question about AI and management. If you are looking at the productivity gains of augmentation of AI agents into your role, then it shouldn’t take long for you to realise that actually managing that productivity is prime for AI takeover!
Everyone wants to know: “Will AI replace me?”
But here’s the question that should keep us all awake at night:
“Will AI manage me?”
Because both are upon us right now!
How We Got Here
Let me tell you a story about how we accidentally designed managers out of part of the job we claimed they should do. For decades we have seen leaders and managers in all sectors and all business devolving “people stuff” to HR.
Manager wants to give feedback? Send them to an HR training course.
Employee has a development need? HR will create a competency framework.
Team member struggling? HR has an Employee Assistance Programme for that.
Performance conversation needed? Here’s an HR-designed nine-box grid and a rating scale.
Slowly, systematically, we stripped managers of the people part of people management.
What did we leave them with?
Performance management.
Tracking. Evaluating. Reporting. Monitoring outputs. Measuring productivity. Ensuring delivery – all answers to the question “What do Managers do”.
And here’s the uncomfortable truth:
AI is really, really good at that.
The Manageable Parts of Management
Think about what we’ve reduced many managers to doing:
- Monitor team productivity metrics
- Track project completion rates
- Review time allocation across tasks
- Identify bottlenecks and blockers
- Allocate resources to priorities
- Flag performance outliers (high and low)
- Generate reports for senior leadership
- Ensure compliance with policies and processes
Read that list again.
Now tell me which of those tasks requires a human.
I’ll wait.
The truth is, AI can do all of that better than most human managers currently do it.
Faster. More objectively. With less bias. At scale. 24/7.
An AI manager doesn’t have recency bias in performance reviews. It doesn’t have favourites. It doesn’t forget to follow up. It doesn’t get overwhelmed when managing 47 direct reports across three time zones. It doesn’t have the kids to pick up!
So if management has become primarily about tracking and optimizing performance, why do we need human managers at all?
The Three Futures Taking Shape
We’re heading toward three possible scenarios. And they’re not mutually exclusive – they might all happen simultaneously in different parts of the same organization.
Future 1: Leaders Leading AI Managers
In this scenario, human leaders don’t directly manage people anymore.
They manage AI agents that manage people.
You’re a Director with 200 people in your organization. But you don’t manage those 200 people – you manage 8 AI managers, each overseeing 25 humans.
The AI managers:
- Track individual and team performance
- Allocate work based on capacity and capability
- Provide real-time feedback on task completion
- Identify skill gaps and recommend training
- Flag issues requiring human intervention
- Generate performance data for review cycles
You, the human leader, focus on:
- Strategy and direction
- Culture and values
- Complex people issues the AI escalates
- Cross-functional relationships
- Innovation and transformation
Sounds efficient, right?
Maybe even liberating – finally free from the administrative burden of people management to focus on “real leadership.”
But here’s the problem:
If you’re not in the daily rhythm of managing people, do you actually understand what they’re experiencing?
Can you lead culture if you’re abstracted from the humans living it?
Can you make good strategic decisions if your only insight into team capability comes from AI-generated reports?
Future 2: Humans Micro-Managed Fully by AI
In this scenario, you show up to work and your manager is an algorithm.
Not metaphorically. Literally.
Your AI manager:
- Assigns your daily tasks based on priority algorithms
- Monitors your productivity in real-time
- Provides immediate feedback on work quality
- Schedules your breaks for optimal performance
- Recommends learning based on skill gap analysis
- Conducts your performance reviews
- Decides your bonus allocation
This isn’t science fiction. Versions of this already exist:
- Amazon warehouse workers managed by algorithms that track their every movement.
- Uber drivers whose earnings and opportunities are determined by AI rating systems.
- Customer service agents whose performance is monitored and evaluated by AI supervisors.
- Gig workers whose next job depends on algorithmic assessment of their last one.
The question is: Does this scale beyond operational roles into knowledge work?
Could your software developer, marketing manager or HRBP be managed by AI?
Could you?
Future 3: The Hybrid Model
This is probably where most organizations will land, at least initially.
Human managers supported by AI co-pilots that handle the “managing” parts of management.
Your manager is human. But they’re heavily augmented by AI that:
- Captures notes on your 121 meetings, then summarises content and context (using the camera to track body language queues or reactions to offer advice on wellbeing or feeling*)
- Pre-drafts your performance review based on data
- Suggests when you need coaching based on pattern recognition
- Recommends task allocation across the team
- Identifies when you’re at burnout risk
- Generates development plans aligned with your career goals
- Can even coach or mentor someone through an avatar of you (again, exists now)
(*I haven’t made this up, it exists right now and you can get full psych reports on the person you are meeting)
The human manager reviews these AI recommendations and makes final decisions.
They add context. Judgment. Empathy. Understanding of circumstances the data doesn’t capture (yet!).
This sounds like the best of both worlds, but it might actually be the most dangerous. It looks human, but it’s increasingly algorithmic.
Your manager thinks they’re making decisions but really they’re rubber-stamping AI recommendations.
Because the AI is so good. So thorough. So data-driven. Why would you override it?
The AI says Sarah is underperforming based on her completion rates. The human manager doesn’t see the context – that Sarah has been mentoring three junior team members, which doesn’t show up in her individual productivity metrics.
The AI recommends promoting James because his output is 23% above team average. The human manager doesn’t notice that James achieves this by never helping colleagues and creating a competitive rather than collaborative team dynamic.
The AI suggests putting Michael on a performance improvement plan. The human manager doesn’t realize Michael’s “decline” in productivity coincides with caring for a sick parent.
In the hybrid model, managers become AI-influenced without realizing how much their judgment has been automated.
It’s not that the AI is wrong, necessarily.
It’s that the AI optimizes for what’s measurable.
And the most important parts of human work often aren’t.
Before you go, what do you think? I am predicting AI agent managers within 3years – are you concerned, disagree or have other ideas let me know in the comments below?
Martin Knowles is an Executive Director of Talent Development with 25+ years leading global learning and talent strategies. He’s currently exploring the uncomfortable questions about AI’s role in management – the ones we need to ask before the future chooses for us.
#FutureOfWork #AIinHR #Management #Leadership #TalentDevelopment #ArtificialIntelligence #HumanResources #WorkplaceCulture #EthicalAI
Responses