Turns out, managing bots is just as annoying as managing people

Sumaiya Noor, chief product officer of a UK-based social-impact investment platform, discusses the challenges of managing AI agents in the workplace. These agents, designed to assist with tasks, can sometimes provide inaccurate answers with great confidence, leading to dissatisfied customers. Noor must periodically refresh the agents on their roles and train them on edge cases, as their performance often deteriorates over time rather than improving.

This experience is shared by other tech leaders, who have found that AI agents are not the efficient solution they once promised. In fact, managing these bots can be just as frustrating as managing human employees. Issues such as "rot," where agents become increasingly less effective over time, long-running tasks derailed by minor errors, and agents duplicating one another's work have been reported.

While AI agents can offload tasks, they do not necessarily provide more free time for employees. Instead, they can lead to mental fatigue, known as "AI brain fry," characterized by mental fog, trouble focusing, sluggish decision-making, and headaches. Workers are also held responsible for any mistakes made by the agents, leading to additional time spent on babysitting and policing legal liability.

The setup and review of AI agents can also pose legal challenges, with compliance risks increasing as the agents become more complex. Despite these challenges, AI agents are becoming more prevalent in the workplace, with 30% of respondents in a Boston Consulting Group report integrating AI agents into their workflows. However, the promise of these agents delivering less work and more efficiency may be overstated, as they can often demand more attention and oversight, leading to increased mental and operational strain for workers.