AI is supposed to make employees more productive, but for managers it is making their jobs harder. A Nielsen survey of 1,700 directors and senior managers across seven countries, commissioned by Monday.com and reported by TechCrunch, found that more than half of senior leaders now struggle to evaluate employees' work. Polished output arriving in inboxes makes it difficult to tell how much was generated by a person and how much by AI. Nearly half of managers said they spend more time reviewing and guiding their teams than before.
Managers are responding by changing the yardstick. Seventy-eight percent say they have revised how they judge employees, putting more weight on skills like crafting prompts, deciding which tasks to delegate to AI, and experimenting with new workflows. Yet that raises fairness concerns: 46% of managers use AI to inform decisions, and 39% use it to draft performance reviews. Leaning on LLMs for reviews may introduce bias, TechCrunch notes; one early-2025 research paper found AI models showed bias against black male candidates, while research from the Luxembourg Institute of Socio-Economic Research found LLMs tended to rate performance similarly to humans when no clear objective standards existed.
AI is also increasing operational complexity. Fifty-seven percent of managers say their organizations now run more tools than before, and only 29% describe their AI infrastructure as fully optimized and centralized. Monday.com's Michal Miller Levi said companies expected AI to shrink software stacks but instead ended up with more apps and integrations. She attributed the broader problem to AI moving faster than organizational processes.
Looking ahead, managers expect to spend more time supervising machines: 58% anticipate overseeing more AI agents than human employees within a year, and 46% worry about losing control over agents. The practical result, the report suggests, is a new layer of invisible labor and a governance challenge as much as a productivity opportunity. With only one source, there is no second set of findings to compare; the survey should be read as a single snapshot rather than a consensus view.