While businesses equip their workforce with AI tools to enhance output, these same tools are emerging as new hurdles for management.
A survey conducted by Nielsen for Monday.com, involving 1,700 senior leaders and directors across seven nations, revealed that over 50% of these managers report increased difficulty in assessing employee performance post-AI integration. This challenge stems from the fact that much of the work submitted by employees now appears highly refined, making it difficult to discern the individual’s contribution versus that of AI. Consequently, almost half of all managers are now dedicating more time to scrutinizing and directing their teams’ work.
Despite these issues, managers largely express positive sentiments towards AI; indeed, 94% of those surveyed expressed enthusiasm for leveraging AI to enhance their work processes, with many confident in identifying its most valuable applications. Michal Miller Levi, Monday.com’s head of market research and customer insights, likened this excitement to the dawn of the internet in the 2000s, noting that people perceive a comparable “magic” in AI.
However, as Levi explained to TechCrunch, the core issue is AI’s rapid evolution, which outpaces corporate adaptation. She elaborated that while AI greatly benefits managers on an individual level, their broader managerial responsibilities—such as organizational leadership—are hindered by AI’s faster progression compared to established organizational procedures.
Furthermore, managers are rethinking their team evaluation methods: 78% of participants indicated a shift in employee assessment post-AI adoption, with increased emphasis on employees’ proficiency in prompt engineering, their ability to judiciously assign tasks to AI, and their willingness to explore novel workflows.
This raises a crucial question: how can employee performance be assessed impartially when managers struggle to distinguish between human and AI-generated contributions? This concern is compounded by the fact that managers themselves are increasingly leveraging AI for evaluation; the survey reported that 46% use AI for decision-making support, and 39% of leaders utilize it for drafting performance appraisals.
Such reliance on AI in reviews risks amplifying existing biases, as Large Language Models (LLMs) fundamentally operate as predictive systems, often replicating human biases present in their training datasets. A study by the Luxembourg Institute of Socio-Economic Research earlier this year indicated that LLMs, when lacking objective standards and evaluating individuals independently, tend to mirror human performance ratings. Moreover, a research paper from early 2025 highlighted that AI models can exhibit significant biases against specific demographics, such as Black male candidates, with biases varying in direction and intensity across diverse social groups.
When questioned about the potential for AI to inject external biases into performance evaluations, Levi acknowledged that companies are still grappling with how to address these challenges.
AI is also transforming the daily responsibilities of managers. It necessitates someone to determine which tasks to assign to AI agents, verify their output, and intervene when errors occur. The report implies that this oversight could constitute a novel form of “invisible labor.”
Managing the software ecosystem itself is also becoming more intricate. Fifty-seven percent of managers reported an increase in the number of tools within their organizations, while approximately 51% noted that their tools have become more interconnected and complex since AI adoption. A mere 29% characterized their AI infrastructure as fully optimized and consolidated.
Levi observed that businesses anticipated AI would lead to a reduction in their software inventories. Conversely, they are now contending with an expanded array of applications, integrations, and the proliferation of ‘shadow AI’—employees utilizing tools not sanctioned by their companies.
Levi stated, “Managers concluded 2025 with the expectation of tool consolidation and a reduction in software count. However, the reality is that the number of tools they employ has increased. Over 50% report using more tools now than last year, and this trend shows no signs of abating.”
The imperative for control is intensifying as companies roll out autonomous AI agents. The survey indicated that 58% of managers anticipate supervising more AI agents than human personnel within the coming year, and 46% expressed concern over potentially losing command of these agents’ activities.
A significant apprehension also exists regarding accountability for AI-generated decisions, cited by 35% of respondents, largely due to the inherent unpredictability of LLMs. Levi clarified that managers prefer human involvement and consultation by AI for critical tasks and decisions, rather than fully autonomous agents.
Essentially, the report’s conclusions highlight that integrating AI is evolving into a significant governance challenge, alongside its potential to boost productivity. Lacking adequate processes and infrastructure to manage AI, managers appear to be struggling to maintain operational efficiency.