Although artificial intelligence holds the promise of boosting workplace efficiency, its integration is presenting unforeseen complexities for managers.
A recent survey conducted by Nielsen, on behalf of Monday.com, involving 1,700 directors and senior managers across seven countries, found that over half of these leaders are now facing greater difficulty in assessing employee performance since the adoption of AI tools. This challenge stems from the fact that AI-generated content often appears highly polished, blurring the lines between an employee's original contribution and AI's assistance. Consequently, nearly half of managers report spending more time reviewing and guiding their teams' work.
Despite these difficulties, managerial enthusiasm for AI remains high, with 94% expressing optimism about its potential to improve work processes. Many also feel confident in identifying areas where AI can add significant value. Michal Miller Levi, head of market research and customer insights at Monday.com, likened this excitement to the early days of the internet, describing AI as possessing a similar 'magic.' However, Levi noted that the core issue is AI's rapid advancement outpacing organizational capabilities to adapt. 'AI performs exceptionally well for individuals, but managers have an additional responsibility to steer the organization as a whole,' Levi explained to TechCrunch. 'Their predicament arises from AI's swift evolution, which is outstripping established organizational frameworks.'
Managers are actively rethinking their approach to employee assessment, with 78% having modified their evaluation criteria post-AI adoption. Increased emphasis is now placed on employees' ability to formulate effective prompts, strategically delegate tasks to AI, and embrace innovative work methodologies. However, a significant concern arises regarding the fairness of evaluations when managers cannot easily distinguish between human and AI contributions. This issue is compounded by the fact that 46% of managers utilize AI for decision-making, and 39% employ it for drafting performance reviews. This practice risks embedding new biases into the review process, as large language models (LLMs) are known to replicate human patterns found in their training data. Research from the Luxembourg Institute of Socio-Economic Research indicates that LLMs tend to mirror human evaluation biases in the absence of clear objective standards. Furthermore, a 2025 research paper highlighted that AI models can exhibit significant biases against certain demographic groups, such as Black male candidates.
AI's influence extends to reshaping a manager's daily routine, introducing a new layer of 'invisible labor' dedicated to supervising AI agents, reviewing their output, and intervening when errors occur. The software landscape itself is becoming more complex; 57% of managers report an increase in the number of tools used, and 51% observe greater interdependence among these tools. Only 29% describe their AI infrastructure as fully optimized. Businesses had initially anticipated AI would streamline their software stacks, but instead, they are encountering a proliferation of applications and shadow AI usage. The growing autonomy of AI agents is also a source of concern, with 58% of managers expecting to oversee more AI agents than human employees within the next year, and 46% fearing a loss of control over their actions. A substantial 35% of respondents expressed apprehension about being held accountable for AI-driven decisions due to the inherent unpredictability of LLMs. Levi emphasized that managers prefer human involvement and consultation by AI for critical tasks, rather than complete autonomy.
Ultimately, the findings underscore that AI adoption is evolving into a governance challenge alongside its promise of productivity. Without robust processes and infrastructure, managers are navigating a complex landscape, striving to maintain operational efficiency amidst rapid technological change.
