In January 2024, Klarna announced that its AI assistant had done the job700 employeescustomer service. A year later, CEO Sebastian Siemiatkowski admitted that the company had reduced its total workforce bythanks to AI 40%— from 5,000 to 3,000 people (CNBC). Microsoft canceled overin 2025 15,000 positionsand openly declared the strategy of „flattening the organization“ — systematic removal of middle management layers (RostNeo). Amazon has automated coordination roles. Meta reduced managers. Shopify has mandated that you have to prove AI can’t handle it before you hire a human (CNBC). TopicAI agents managementis key today.
Key Findings
- Ai agentschanges the rules of the game — companies must act now
- Data from McKinsey and Gartner confirm: early adopters grow 2-3x faster
- The key is to start with a pilot, not a big transformation
- Slovak companies lag behind by 2-3 years — the window of opportunity is closing
- Investment in AI returns within 18 months if deployed correctly
That’s not the future. This has happened in the last 18 months.
The question for you is not,whetherit’s happening. The question is: what percentage of your organization is at risk today — without even knowing it?
Chapter 1: The Data Middle Management Doesn’t Want to Hear
Let’s start with the inconvenient truth. The World Economic Forum in itsFuture of Jobs Report 2025states that22% of all jobswill undergo a fundamental transformation by 2030. 170 million new roles will be created, but 92 million will disappear — although the net difference is positive (+78 million), but the distribution is brutally uneven (WEF).
Who will disappear first? Not workers at the belts. Not creatives.Coordinators. Reporters. Information intermediaries.Exactly the roles that most companies call „middle management.“
McKinsey in itsState of AI 2025the report states that23% of organizations are already integrating agent AI systemsacross the enterprise and another 39% are experimenting (McKinsey). That means nearly two-thirds of large companies are actively testing or deploying AI agents — autonomous systems that don’t need a human intermediary to coordinate workflow.
Fakt:Gartner predicts that40% of enterprise applicationswill have integrated AI agents by the end of 2026, up from less than 5% in 2025 (Gartner).
What does this mean for middle management? Think about what your team leaders, project managers and coordinators do all day:
- Gather information from one team and transfer it to another
- Create reports and status updates
- Organize meetings and track deadlines
- Escalate issues and push decisions nadol
Presne totoAI agents do better, faster and 24/7. Not in a year. Now.
AI agents management: Chapter 2: AI agents 2026 — not chatbots, but autonomous collaborators
Just to be clear: when we talk about AI agents in 2026, we’re not talking about a ChatGPT box where you type questions. We are talking about systems that:
- They perform multi-step tasks independently— from problem identification through data analysis to solution implementation
- They communicate with other agentsalso by people as needed
- They make decisions autonomouslywithin defined limits
- They are learning and improvingbased on the results
Gartner predicts that by 2029 it will be70 % podnikovdeploy agent AI as part of the IT infrastructure (Gartner). McKinsey estimates that AI agents and robotics will create2.9 trillion dollarsof annual economic value in the US by 2030.
Why is this relevant to middle management?
Because an AI agent does exactly what a coordinator does: collects inputs, evaluates them, escalates exceptions, and implements decisions. But the agent does it without breaks, without political games, without information silos and without “I forgot to forward you that email.”
Shopify got it.CEO Tobi Lütke ordered in April 2025: “Reflexive use of AI is now abaseline expectation.in Shopify.“ Every manager has to prove that AI can’t do it before asking for a new human (MIT CDO). The use of AI has become part of the performance review.
This is not a memo from an enthusiastic tech CEO. That’s the new standard.
Chapter 3: Slovak reality — a silent bubble of coordination roles
In Slovakia, we have a specific problem:historically increased layers of middle management, which arose not from an operational need, but from a culture of control and mistrust.
A typical Slovak company with 200 employees has 4-6 levels of management. Compare this to Nordic firms of the same size where there are 2-3 levels. Each extra layer means:
- Slower decision-making (information has to go through more filters)
- Higher costs (managers are more expensive than specialists)
- Less engagement (people far from decisions lose motivation)
The PwC CEE CEO Survey repeatedly shows that Slovak companies lag behind in digital transformation compared to Western and Baltic countries. The reason is not in the technology — you can buy it.The reason is in the organizational structure, which was designed for an era when information was scarce and distribution required people.
Today, the information is distributed by the API. Workflow is coordinated by an agent. The status report is generated by the dashboard. What is left for the middle management?
Slovak caution— „we’ll wait to see how it turns out for others“ — is a strategic trap in this case. While you wait, your competition in the Czech Republic, Poland and Hungary reduces layers and speeds up decision-making. The talent you want to retain is leaving for flat organizations where they have autonomy and direct impact.
Chapter 4: What Elite Firms Do Differently — From Hierarchy to Agency Organization
McKinsey in its2026 State of Organizationsreport confirms a shift to what it calls„agentic enterprise“— an organization designed around AI agents, not management layers (ReplyFabric).
What does it look like in practice?
Nvidia — „Mission Teams“:Jensen Huang runs one of the most valuable companies in the world with an extremely flat structure. Instead of management layers, he has cross-functional “mission teams” with a direct line to leadership. AI tools ensure coordination and reporting. The result: decision-making speed that hierarchical organizations cannot replicate.
Microsoft — strategic “flattening”:In 2025, Microsoft openly declared a strategy of removing the middle layers. Goal: “widening the span of control” — each manager leads more people directly, as AI tools take over the coordination burden (The HR Digest). It’s not a savings — it’s an architectural change.
Shopify — „AI-first headcount“:Before each team increase, the manager must demonstrate that The AI solution was not enough. AI becomes first choice, human second — not because people are not important, but because coordination roles (not human roles) are ineffective (Fast Company).
The common denominator:Elite companies are not eliminating people.They are removing layers.They are moving people closer to customers, to decisions, to creating value. AI is taking over what people did only because there was no better way — distribution of information, coordination and reporting.
Chapter 5: Three Scenarios for the Middle Manager — Adapt, Transform, or Exit
If you are a middle manager, this is not a reason to panic. It’s a cause for action. You have three paths:
Scenario A: Adapt — become an “AI-augmented manager”
Adopt AI agents as part of your team. Instead of spending 60% of your time coordinating and reporting, delegate it to AI and focus on:
- Coaching— people development, feedback, mentoring
- Rozhodovanie v neistote— where data is not enough and judgment is needed
- Stakeholder navigation— relations, politics, negotiation
WEF data confirms: the fastest growing competencies areanalytical thinking, creativity, resilience and curiosity (WEF). Not coordination. Not reporting.
Scenario B: Transform into a different role
.Project Manager →
- .Team Lead →Product Owner(ownership of the result, not the process)
- Team Lead → People Developer(people, not tasking)
- Report Manager → Decision Scientist(interpretation, not data collection)
- Operations Coordinator → AI Operations Designer(you design how AI works)
Scenario C: Leave — and start where the future is
For some, the best choice is to leave for an environment where their skills will be more valuable. AI-native startups, consulting, education — all these fields need people who understand both people and technology.
The worst choice?Do nothing and hope it doesn’t affect you. According to Gartner, the atrophy of critical thinking caused by passive reliance on GenAI will force50% of organizationsrequire „AI-free“ competency assessments (Gartner). The market will reward those who AI activelymanages, not those who are passivelyreplaced by.
Chapter 6: The Klarna case — what happens when a company gets it wrong
Klarna is a perfect case study of that,how not to doAI transformation of middle management.
In 2024, the firm announced with fanfare that AI had replaced 700 people. The media celebrated it. The CEO spoke at conferences. Stocks rose.
Two years later, the reality: customer satisfaction has fallen on complex interactions. Edge-cases, emotionally demanding situations and multi-level problems were not handled by AI.55% of companies that made AI-driven layoffs regret it today (Digital Applied). Klarna had to start a rehiring program.
What went wrong?
- Klarna replacedpeople, instead of replacinglayers and processes
- It did not retain human-in-the-loop for complex scenarios
- She didn’t retrain existing people — she simply fired them
- She underestimated the emotional and empathetic component of customer service
Lessons for leaders:Transforming the organizational structure is not the same thing as mass layoffs. The goal isn’t „fewer people“ — the goal is „people in the right places, AI in the right places.“
Chapter 7: Leading by Example—Your Responsibility to the People You Hire
As FCH, we believe that transformation without humanity is not transformation — it is destruction.
If you are a CEO or director and you know that AI agents will gradually take over coordinating roles in your company, you havemoral and strategic responsibilityprepare your people for itbeforebefore it happens. Not after.
„We solve ahead“— this is the principle that distinguishes elite companies from those that react only in a crisis.
Specific steps:
1. Layer Audit (this month):Count how many layers you have between the CEO and the customer. If it’s more than 4 — you have a problem. Identify roles whose primary function ismoving information, notvalue creation.
2. Transparent communication (this quarter):Tell your people the truth. Not „AI is not replacing us“ (because that’s not the whole truth). You say, „Sometasksare going to change. We want to prepare you. Here’s the plan.“ Trust is built with truth, not PR phrases.
3. Reskilling program (6 months):Create a transition plan for any manager whose role is primarily coordination. Where can it move? What does he have to learn? Who will mentor him?41% of organizations plan to reduce their workforcein rolls with AI exposure — but70% plan to hire people with AI skills (WEF). That means: same people, new skills.
4. Pilot project of the agency structure (Q3 2026):Select one team. Deploy an AI agent for coordination, reporting and status tracking. Redirect the manager’s free time to coaching, development and strategic decision-making. Measure results. Iterate.
In 2030, we will look at today’s middle management the way we look at the telephone operators of 1960. Not with disdain — with the understanding that technology has moved the world on.
The difference between the leaders who todaythey remember the namesof their people, and by those who remember onlyheadcount, will define what companies will be here in five years.
Your people are not rows in Excel. They are your responsibility.
Transform them before the market transforms them — without your guidance.
Frequently asked questions
What does AI agents mean for Slovak companies?
Ai agents is a key topic for Slovak companies in 2026. The article analyzes specific data, trends and recommendations based on McKinsey, BCG and Gartner research. Leaders must act now to maintain a competitive edge.
How to implement AI agents in practice?
Implementing AI agents requires a strategic approach — first an audit of the current state, then a pilot project and gradual scaling. The key is to involve the company’s management and build internal expertise.
What is the outlook for AI agents by 2027?
Trends show that AI agents will be an increasingly important topic. According to WEF and Gartner, the adoption of AI is expected to accelerate, regulations will tighten and the pressure for data-driven decision-making will increase. Companies that start acting now will get a 2-3 year head start.
