AI Is Speeding Up Work — and Slowing Managers Down, New Survey Finds

Artificial intelligence is helping employees turn out cleaner, faster work, but a new international survey suggests it is also making management harder. More than half of senior leaders say AI has made it more difficult to evaluate employee performance, while many now spend extra time checking, guiding and reworking the output that lands in their inboxes.

The findings, released on October 21, 2026, come from a Nielsen survey of 1,700 directors and senior managers across seven countries and were commissioned by work-management software company Monday.com. The results point to a growing gap between AI’s promise of higher productivity and the practical challenge of managing teams, workflows and accountability in organizations that are adopting the technology faster than they can adjust their processes.

Why managers say AI has become a headache

AI is supposed to save time. For managers, however, it is increasingly creating a new layer of review work that did not exist before. The core complaint in the survey is simple: when employee work arrives polished and presentation-ready, managers struggle to tell how much of it reflects human effort and how much came from a chatbot, writing assistant or other AI tool.

That uncertainty has consequences. More than half of respondents said AI has made performance evaluation harder, and nearly half said they now spend more time than before reviewing or coaching the work their teams produce. In other words, the software may be speeding up delivery at the bottom of the org chart, but it is also adding friction at the top.

What changed after AI arrived?

According to the report, managers are no longer judging only the final outcome. They are also judging how employees get there, and that shift is reshaping daily oversight. The survey found that 78% of managers have changed the way they assess employees after AI adoption, putting more emphasis on skills such as prompt design, task delegation to AI and experimentation with new workflows.

That means managers are increasingly expected to reward not just polished deliverables, but also the judgment behind them. The rise of AI has turned technical fluency into part of performance evaluation, even as companies are still debating what “good” AI use looks like in practice.

How managers are redefining performance reviews

Managers are adapting by changing what they look for. Instead of focusing solely on traditional output metrics, many are now asking whether employees can use AI responsibly, efficiently and creatively.

The survey suggests that prompt-writing ability, tool selection and workflow experimentation are becoming part of the new managerial vocabulary. For many companies, that is a major shift: the skill set being measured is no longer limited to subject-matter expertise or speed, but also includes the ability to collaborate with machines.

Michal Miller Levi, head of market research and customer insights at Monday.com, said the current excitement around AI resembles the early internet era, when new technology seemed to carry an almost magical sense of possibility. But she added that companies are struggling because AI is advancing faster than internal processes can be redesigned to absorb it.

Levi told TechCrunch that AI can work well for managers as individuals, but their responsibilities extend beyond their own productivity. They must also help steer teams and organizations, and that broader role is where the strain begins. In her view, the technology is moving more quickly than companies can update governance, workflows and expectations.

Who is deciding whether AI should help with reviews?

Managers themselves are increasingly leaning on AI to shape those judgments. The survey found that 46% use AI to help inform decisions, while 39% say they use it to draft performance reviews.

That creates an awkward feedback loop: employees may be judged partly on work touched by AI, while managers may be using AI to assess that work. The result is a process that can become faster, but not necessarily fairer or more transparent.

Could AI make performance reviews more biased?

Yes, AI could add new layers of bias, especially if companies adopt it without clear standards. The survey’s findings raise concerns because large language models are trained on human-generated patterns and can reproduce those patterns in their recommendations and wording.

Research cited in the report points in the same direction. A study from the Luxembourg Institute of Socio-Economic Research earlier this year found that when objective standards are missing and people are rated independently, large language models tend to score performance similarly to humans. Another early-2025 research paper found that AI models can show measurable bias against Black male candidates, with the direction and strength of those biases varying across demographic groups.

Asked whether AI-generated input could bring outside bias into performance management, Levi said companies have not yet settled on how to handle the issue. That uncertainty is important: the more AI enters decision-making, the more pressure there is to define who is accountable for the result.

The hidden labor behind AI supervision

One of the report’s strongest themes is that AI does not eliminate management work so much as rearrange it. Someone still has to choose which tasks are appropriate for automation, inspect the results, and step in when a system goes wrong or produces something off target.

The survey describes this as a new form of invisible labor. Managers are not just managing people anymore; they are also managing machine output, deciding when to trust it and when to intervene. That supervision takes time, even if the final output appears to have been generated instantly.

This dynamic helps explain why many leaders are not simply celebrating AI productivity gains. They are also absorbing new responsibilities that are harder to see on a balance sheet but still matter for day-to-day operations.

How AI is changing the software stack

The adoption of AI has also made the workplace software environment more complicated, not less. The survey found that 57% of managers say their organizations now use more tools than before, while 51% say those tools have become more interdependent and clustered since AI entered the mix.

Only 29% described their AI infrastructure as fully optimized and centralized. That is a striking figure, because many companies assumed AI would help consolidate platforms and streamline workflows. Instead, many appear to be adding apps, integrations and shadow tools as employees experiment outside approved systems.

Levi said businesses expected AI to reduce the size of their software stacks. Instead, they are dealing with more fragmentation, more connections to maintain and more unsanctioned usage by employees who turn to outside tools when official ones do not meet their needs.

What do managers fear most about AI agents?

Managers are not only dealing with better AI-generated documents; they are also preparing for systems that can act with increasing autonomy. The survey found that 58% of managers expect to oversee more AI agents than human employees within the next year, a dramatic signal of how quickly expectations are shifting.

At the same time, 46% said they worry about losing control over those agents’ actions, and 35% said they are concerned about being held responsible for decisions made by AI. Those are not abstract worries. They go to the heart of workplace accountability: if a machine takes an action that causes a problem, who owns the outcome?

Levi said leaders do not want fully autonomous agents making important choices on their own. They want people involved in key decisions, especially when the stakes are high. That preference underscores a broader truth about enterprise AI: companies may want speed, but they are still uncomfortable surrendering control.

How the survey reframes the AI productivity debate

The Monday.com survey adds nuance to the familiar story that AI boosts efficiency. That story is not wrong, but it is incomplete. Productivity gains at the employee level can create governance problems at the managerial level, especially when organizations lack policies for review, auditing and escalation.

In that sense, AI adoption is no longer just a technology decision. It is a management design problem. Companies must decide how to measure work, who reviews machine-assisted output, how to write performance standards and how to centralize tools without stifling experimentation.

The report suggests many organizations are not there yet. Instead, they are improvising in real time, trying to capture the benefits of AI while minimizing the confusion it introduces. That is especially difficult when the technology evolves faster than internal policy cycles, and when employees and managers are both using AI in ways that are not always visible to the company.

Why organizational processes are lagging

Companies are moving faster on AI adoption than on management reform. That mismatch helps explain why the same technology can feel liberating to employees and burdensome to supervisors. If teams can produce more polished work in less time, managers must still build new systems to evaluate it, approve it and maintain quality control.

In practice, that means AI may be shifting labor rather than removing it. Writing, summarizing and drafting can happen more quickly, but oversight, verification and decision-making still require humans. The manager’s role becomes less about producing the first draft and more about validating the final one.

Timeline: How this management problem emerged

The shift did not happen overnight. It has unfolded alongside the broader rollout of generative AI into daily office work.

What the findings mean for employers

For employers, the survey is a warning that AI rollout should not be measured only by employee adoption rates or productivity metrics. If managers are spending more time correcting output, evaluating prompt use and handling tool sprawl, the company may be trading one kind of labor for another.

That does not mean AI has failed. It means organizations need better operating rules. Clear standards for review, human oversight for sensitive decisions, and more disciplined software governance may matter just as much as the tools themselves.

Companies may also need to rethink what counts as a strong employee. In an AI-enabled workplace, effectiveness may increasingly include the ability to work with machine assistance while still demonstrating originality, judgment and accountability.

The bigger picture

The broader lesson from the survey is that AI is becoming as much a governance issue as a productivity issue. The more organizations rely on AI systems, the more they must answer difficult questions about trust, responsibility and control.

That tension is likely to intensify as AI agents become more common. If the technology is used to draft reports, suggest decisions and complete tasks with limited supervision, managers will need new ways to monitor work without slowing it to a crawl. They will also need to protect against bias, error and unclear accountability.

For now, the survey indicates that many managers are enthusiastic but overstretched. They see the promise of AI clearly, but they are also discovering that the path to real productivity gains is paved with extra review, more policy work and a larger burden of oversight.