Opinion
The battle for managerial autonomy: Are your employees managing AI, or is AI managing them?
"The question is no longer technical but fundamentally managerial: if the machine knows how to execute, who defines what is worth executing?" writes Maya Blanc, VP of HR at Evinced.
The economic headlines of the first quarter of 2026 leave no room for interpretation. We are in the midst of a classic productivity paradox: investments in Artificial Intelligence are at an all-time high, yet results remain mediocre at best. The question is, why hasn't the promised ROI materialized? It is not because the algorithms aren't good enough; rather, it is because organizations are still attempting to manage cutting-edge technology with outdated mindsets.
The automatic tendency is to look for the problem in the code, the models, or the infrastructure. But in practice, the bottleneck is located in a much less comfortable place: the boardroom. It’s in the managerial culture and in the way we define control.
The promise of AI has already shifted. If a year ago we were talking about tools that generate content, today we are deep into the era of Agentic AI- systems that don’t just react, but act. They break down business goals, make intermediate decisions, and execute. In such a reality, the question is no longer technical but fundamentally managerial: if the machine knows how to execute, who defines what is worth executing?
Management in a State of Emergency
Israel's security reality has only sharpened this dilemma. The war, and specifically the escalation with Iran, didn't create a new problem; it exposed an existing weakness. The classic management model—based on visibility, oversight, and tight coordination—simply doesn't hold up when people are working between sirens, bomb shelters, and a total lack of routine.
In this situation, a manager has no real way to "manage work" in the traditional sense. There is no control over hours, pace, or sometimes even availability. What remains is only one thing: the extent to which the employee can manage themselves. This is where a concept transitions from a theoretical recommendation to a survival criterion: Agency.
Agency is not just "taking initiative" or another variation of "empowerment." It is a professional identity perception: to what extent do people see themselves as catalysts for change rather than merely as reactors to reality? It’s the difference between an employee waiting for a task and one who defines it; between the one who executes and the one who decides what needs to be done in the first place.
The Human Advantage: Judgment Over Execution
In the AI era, this value becomes critical. When technology can already write code, build architectures, and suggest solutions, the human advantage shifts. It is no longer found in the speed of execution, but in the quality of judgment- the ability to identify the right problem, not just the fastest solution. Above all, it lies in taking responsibility for the outcome, not just the action.
Here lies the contradiction that many managers still ignore: we are giving AI greater autonomy, yet we continue to manage people as if they were operational components.
The result is predictable. AI progresses rapidly but gets stuck at decision points. And the employees? They burn out. Instead of becoming process leaders, they become machine operators. This leads to less responsibility, less meaning, and in the long run, less commitment.
Organizations that successfully break this paradox understand that Agency is not a "nice-to-have "; it’s a prerequisite. This applies not only to employee development but to the hiring process itself. In an environment of uncertainty, the ability to act without precise instructions is not a personality trait; it is a professional skill.
It is also important to stop thinking of Agency as a binary trait. It’s not "either you have it, or you don’t"; it’s a scale. There are employees who execute, those who solve problems, and a rare few who redefine the problem itself and the value the organization produces.
The ability to identify these levels and build growth paths accordingly is no longer a privilege of "advanced HR." It is a critical managerial tool, especially in a period where external pressure is high and internal meaning becomes a retention engine no less important than salary.
But this change doesn't happen on its own. It requires letting go—and that’s the unpopular part. Managers must relinquish the illusion of control and move from supervisors to coaches. They must define the direction but not micro-manage every step. Most importantly, they must allow employees to make real decisions, including the risk of mistakes.
At the same time, employees cannot remain passive and wait to be "allowed." In a world where machines are becoming increasingly proactive, remaining reactive is a dangerous choice. Those who wait for instructions will find themselves working for the AI. Those who take ownership will manage it.
Ultimately, amidst all the technological and security noise, this is perhaps the most practical question every organization must ask itself today: Are we building employees who lead systems, or operators who serve them?
Maya Blanc is the VP of HR at Evinced.
