AI May Boost Workplace Productivity, but It Is Giving Managers New Challenges

Table of Contents

• You might want to know
• AI Productivity and the New Management Burden
• Key Insights Table
• Afterwards...

You might want to know

• How can managers fairly evaluate work when AI has contributed to the finished result?
• What new responsibilities arise when companies introduce AI agents and additional tools?

AI Productivity and the New Management Burden

Companies are giving employees artificial intelligence tools in the hope of making work faster and more productive. Yet the tools are creating an unexpected management challenge: when AI helps produce polished work, it can become harder to understand how much of the result reflects an employee’s own effort, judgment, and skill. Instead of simply saving time, AI can add new work for managers who must review, interpret, and guide the output of their teams.

A Nielsen survey of 1,700 directors and senior managers across seven countries, commissioned by work-management company Monday.com, found that more than half of senior leaders say evaluating employees’ work has become harder since adopting AI. The difficulty is not necessarily that the work is poor. In many cases, the opposite may be true: employees can use AI to produce work that appears complete and professional. But the finished product may not make clear which parts came from the employee and which were generated or shaped by AI. Nearly half of managers say they now spend more time reviewing and guiding their teams’ output than they did before.

That added effort exists alongside broad enthusiasm for the technology. In the survey, 94% of managers said they were excited about improving work with AI, and many expressed confidence that they understand where it can deliver the most value. Michal Miller Levi, head of market research and customer insights at Monday.com, compared this enthusiasm to the early days of the internet in the 2000s. In both cases, people have seen a kind of “magic” in the technology and its potential to change how work gets done.

However, excitement does not automatically give organizations the time or systems they need to adapt. Speaking to TechCrunch, Levi said AI is developing faster than organizational processes can keep up. AI may work well for individual managers in their own tasks, she explained, but managers also have a broader responsibility to guide their organizations. Their challenge is therefore not only learning to use AI themselves; it is helping teams and business processes adjust to a rapidly changing technology.

As AI becomes part of everyday work, organizations are also reconsidering how they assess employees. The survey found that 78% of respondents had changed their evaluation methods after adopting AI. Greater weight is now being placed on skills such as writing effective prompts, deciding which tasks to delegate to AI, and experimenting with new ways of working. These capabilities may help employees use AI productively, but they also raise questions about what performance should mean when a task is shared between a person and a machine.

Fairness becomes a particular concern when managers cannot easily identify the source of an employee’s work. A polished result may reflect substantial human expertise, extensive AI assistance, or some combination of both. If evaluation criteria do not account for these differences, employees could be judged inconsistently. Managers may need clearer expectations about which uses of AI are acceptable, what kinds of human contribution matter, and how employees should disclose or document AI assistance.

The process becomes more complicated when managers use AI to help make or communicate decisions about staff. The survey found that 46% of managers use AI to help inform their decisions, while 39% of leaders use AI to draft performance reviews. These applications may save time, but they also introduce risks into a high-stakes process. A generated review can sound authoritative without necessarily being accurate, complete, or appropriate to the individual being assessed. Managers remain responsible for checking whether the language reflects evidence and for ensuring that judgments are based on relevant standards.

There is also a broader concern about bias. Large language models are prediction systems that learn patterns from training data, and those patterns may reproduce human assumptions. Research from the Luxembourg Institute of Socio-Economic Research earlier this year found that, “in the absence of clear objective standards and when individuals are rated independently,” LLMs tend to rate performance similarly to humans. That finding suggests that AI does not automatically remove subjective judgment from evaluations; it may reproduce familiar patterns unless organizations use clear, consistent criteria.

A research paper in early 2025 also found that AI models can be biased against black male candidates and “demonstrate significant biases, varying in terms of the directions and magnitudes across different social groups.” This underscores why human oversight and careful evaluation matter when AI is used in decisions affecting people. Asked whether AI-assisted performance assessments could introduce outside bias, Levi said companies have not yet worked out how to manage such issues. The uncertainty points to a need for organizations to establish safeguards rather than assume the technology will produce neutral judgments.

AI is changing not only how managers evaluate work, but also what their day-to-day jobs involve. Someone must decide which tasks to hand to AI agents, examine what those agents produce, and intervene when results are incorrect or problems arise. That oversight can become a substantial but hard-to-see part of the work. The report describes it as a potential new layer of “invisible labor”: the time and attention required to supervise automated activity may not appear in the final output, even though it is necessary to produce dependable work.

The technology environment itself is also becoming more complicated. Fifty-seven percent of managers say their organizations now have more tools in circulation than before, and about 51% say their tools have become more interdependent and clustered since adopting AI. Only 29% describe their AI infrastructure as fully optimized and centralized. These findings suggest that introducing AI does not necessarily simplify the workplace’s software landscape. It can instead create more connections to manage, more systems to understand, and more potential points of failure.

According to Levi, many businesses expected AI to help them reduce their software stacks. Instead, they are dealing with more applications, integrations, and shadow AI use as employees adopt tools outside company-approved systems. Levi said managers completed 2025 with an expectation that they would consolidate tools and reduce software. In practice, the number of tools increased: “More than 50% are saying that they’re using more tools now than they did last year, and it’s not stopping […].” This gap between expectation and reality makes it harder to maintain a clear view of what systems employees are using and how those systems connect.

These concerns are becoming more urgent as companies deploy AI agents capable of acting with greater independence. The survey found that 58% of managers expect to oversee more AI agents than human employees within the next year. At the same time, 46% worry about losing control over what those agents do. Another 35% of respondents said they fear being held responsible for decisions made by AI, even as large language models remain largely unpredictable.

Managers’ concerns do not necessarily mean that they reject automation. Rather, they appear to want a clear role for people in decisions that carry significant consequences. Levi said managers do not want their agents to be fully autonomous. They want people to be involved and consulted when AI handles crucial tasks and decisions. In practical terms, that means organizations need to decide which actions agents can take independently, which require approval, and who is accountable when an automated process produces an error.

Overall, the survey’s findings suggest that AI adoption is becoming as much a question of governance as an opportunity to raise productivity. Tools can help employees and managers complete work, but their value depends on how responsibly they are integrated into existing processes. Organizations need to establish evaluation standards, clarify expectations for AI-assisted work, review potential bias, maintain oversight of agents, and understand the software systems employees use. Without suitable processes and infrastructure, managers may be left scrambling to keep work moving while trying to control a growing set of tools and decisions.

Key Insights Table

Afterwards...

The next phase of workplace AI will depend not only on what the technology can do, but also on whether organizations can build reliable ways to use it. Clear performance standards, sensible limits on agent autonomy, transparent processes, and consistent human review can help companies benefit from AI without losing sight of accountability or fairness.

For managers, the task ahead is to balance experimentation with oversight. AI may take on more routine work, but supervising its output, integrating tools, and making sound decisions about people will remain important responsibilities. Companies that recognize this management work and provide the systems and guidance to support it will be better positioned to turn AI’s promise into sustainable productivity.