While artificial intelligence promises to eliminate routine tasks, adopting it requires a significant upfront investment of time. Personally, I have eliminated hours of administrative overhead by deploying AI agents to manage repetitive assignments. Yet, for countless professionals, managing these systems has morphed into an exhausting full-time job.

We have all experienced the friction of drafting, erasing, and refining prompts, verifying outputs, and constantly fixing machine-generated errors and flawed assumptions. Recently, researchers from the Glean Work AI Institute introduced a precise label for this hidden operational overhead: botsitting. For a staggering number of employees, this phenomenon has transformed into an acute workplace crisis, depleting their energy, motivation, and the very tasks they once found fulfilling.

The Hidden Cost of Managing AI Agents

Data from a survey of 6,000 white-collar professionals reveals that the typical respondent spends 6.4 hours every week botsitting. That investment nearly matches the time they dedicate to using AI for actual productivity. The core issue is that employees find zero fulfillment in this oversight. Rebecca Hinds, a co-author of the study, noted during a recent podcast appearance that AI management is widely viewed as tedious, exhausting, and completely ignored, tracked, or valued by executive leadership.

Consequently, professionals burdened by botsitting are 73% percent more likely to actively seek new employment. The report issues a stark warning: workers who shoulder this hidden labor without acknowledgment or reward grow exhausted, then resentful, and finally start updating their résumés. This trend demands immediate attention from corporate leadership, particularly since the study indicates that employees leveraging AI most aggressively are also those facing the highest risk of burnout.

Fortunately, losing top talent to automation fatigue is entirely preventable. Burnout fundamentally stems from a combination of overwork and underappreciation. Botsitting-induced exhaustion follows this exact pattern, though it remains notoriously difficult for leaders to recognize work they cannot see.

Conducting an AI Time Audit

To bridge this visibility gap, leaders should implement targeted AI time audits. Ask your teams to track every minute dedicated to AI over a single week. This includes prompt optimization, switching between multiple applications, and rigorously auditing the generated results.

Analyzing this data provides clear visibility into operational bottlenecks. If multiple team members struggle with a specific platform, the software may simply lack the maturity for the assignment. Conversely, if an employee repeatedly runs identical prompts to secure one usable outcome, targeted training can immediately resolve the friction.

Knowing When to Deploy—And Restrain—Artificial Intelligence

At Jotform, team members are continuously encouraged to experiment with AI platforms to streamline workflows. These trials have delivered remarkable wins, ranging from streamlined onboarding procedures to automated customer support triage.

However, plenty of initial experiments failed the reality test. Certain applications generated polished text that demanded so much heavy editing that any potential time savings completely vanished. In other instances, the initial configuration and ongoing supervision consumed more resources than the manual task ever did.

Such friction is an inherent part of innovation. Thomas Edison tested thousands of filament variations before inventing a functional incandescent bulb, and James Dyson engineered more than 5,000 prototypes before perfecting his bagless vacuum. Effective AI experimentation demands the exact same mindset: every failure yields vital insights, even if it simply proves that a particular process should remain human-driven.

Leaders must avoid penalizing teams when an automation initiative falls flat. By formally integrating experimentation into daily workflows, testing—and encountering failure—becomes a standard component of the role rather than a risky individual endeavor.

Safeguarding the Work People Love

During a recent conversation with a writer friend, he shared conflicted feelings regarding artificial intelligence. He acknowledged its undeniable utility for brainstorming and research, yet admitted those exact activities have always been his favorite parts of the craft. When we automate those responsibilities away, what remains of the profession?

New York Times columnist Bret Stephens similarly addressed the dangers of delegating writing to algorithms, observing that outsourcing thought is mentally enfeebling—akin to taking an escalator when daily stairs are required to maintain strength. This principle extends far beyond writing. Any capability left unpracticed inevitably deteriorates, and AI frequently targets the exact skills professionals cherish most. According to the research report, 51% of workers report that AI has already automated meaningful responsibilities they would have preferred to handle personally.

Automation should always serve to protect meaningful labor, never to eradicate it. During Jotform's formative years, I spent entire days answering support tickets, forcing the work I truly valued to the absolute margins. Hiring a support division was never about doing less; it was about reclaiming the craft I loved. Decades later, that philosophy guides every technological decision: distinguishes mundane drudgery from the actual art of the work.

The solution is not to retreat from innovation, but to deploy it with rigorous discernment. Track experimental hours, evaluate what fails, and ensure your people retain the responsibilities that inspire them.