The rapid integration of Artificial Intelligence (AI) into the modern workplace, a transformation largely spearheaded by Human Resources departments, has shifted from a focus on adoption to a more nuanced challenge: ensuring AI is used efficiently and effectively to truly capture its potential. While initial efforts concentrated on getting employees to embrace AI tools, a new report from Glean reveals a significant, yet often overlooked, consequence of widespread AI deployment: "botsitting." This phenomenon, where employees spend a substantial portion of their time supervising, correcting, and essentially managing AI outputs, is not only diminishing the promised productivity gains but is also leading to employee burnout and disengagement.
The Rise of Botsitting: An Emerging Workplace Phenomenon
Recent data indicates that AI adoption is nearly ubiquitous among digital workers. A comprehensive survey of 6,000 professionals, as detailed in Glean’s Work AI Index Report, found that approximately 87% of respondents are actively using AI in their daily tasks. A significant majority, around three-quarters, believe AI enhances their productivity, with many estimating time savings of up to 11 hours per week. However, the stark reality is that only a modest 13% of organizations are experiencing a tangible improvement in overall performance. This disconnect stems from the substantial time employees are dedicating to what Glean terms "botsitting."
Defining Botsitting: The Hidden Labor of AI Integration
Botsitting is defined as the essential, yet often unacknowledged, work required to make AI systems functional and reliable. This includes tasks such as providing missing contextual information, meticulously reviewing AI-generated outputs for accuracy, debugging errors, re-running prompts when initial results are unsatisfactory, and correcting the confident but erroneous responses that AI can sometimes produce. In essence, employees are acting as human intermediaries, bridging the gap between AI’s capabilities and its practical application.
Eva Spatz, Vice President and Head of People Experience at Staffbase, emphasizes that botsitting is fundamentally an HR issue, rooted not in technical flaws of the AI itself, but in organizational dynamics. "It stems from a lack of trust and psychological safety, not a software glitch," Spatz explains. "If employees lack a clear purpose, or worry that an AI mistake could cost them their jobs, they get trapped acting as rigid, hyper-vigilant operators rather than impact architects." This psychological aspect is crucial; when employees feel their roles are at risk due to AI, their approach to using these tools becomes overly cautious and focused on risk mitigation rather than innovation.
Quantifying the Time Sink: The True Cost of AI Oversight
The Glean report provides alarming statistics on the time employees are dedicating to botsitting. On average, workers are spending an estimated 6.4 hours per week on these oversight activities, equating to nearly a full workday lost to managing AI. When analyzing the total time employees spend interacting with AI tools, a striking 37% is consumed by botsitting, closely followed by 36% spent on actual work output generation. This means that for every hour spent leveraging AI for productive tasks, an equivalent amount of time is spent troubleshooting its shortcomings. Furthermore, over a quarter of AI interaction time is now dedicated to learning and developing AI agents, indicating a continuous investment in mastering these evolving technologies.
The nature of botsitting work is often described as "grunt work." This includes repetitive tasks like re-inputting context into different AI applications, identifying and rectifying AI "hallucinations" (fabricated information), and verifying outputs that may sound authoritative but are factually incorrect. Researchers note that AI can sometimes provide answers that employees want to hear, rather than what is truthful, necessitating diligent human verification.
While a small portion of this oversight can be considered productive, it comes at a significant talent cost. This "hidden labor" is often invisible, unbudgeted, and unsupported by organizational structures. Employees who absorb this workload without recognition or reward are prone to exhaustion, resentment, and ultimately, a desire to seek opportunities elsewhere. The Glean findings reinforce this, indicating a strong correlation between increased time spent botsitting and a higher likelihood of employees actively searching for new employment.
The HR Imperative: Building Human Infrastructure for AI Success
The findings underscore the critical role of HR in navigating the complexities of AI integration. Resisting the urge to rush AI transformation without careful consideration of its human impact is paramount. "Organizations must build the human infrastructure—not just the technology infrastructure—that makes AI worth using, or they’ll keep paying the bill," the researchers advise. This means prioritizing the human element in AI strategy, ensuring that technological advancements are supported by robust people-centric practices.
Spatz advocates for people-centric leadership that treats psychological safety as a fundamental component of operational infrastructure. This involves a proactive approach to redefining job roles, emphasizing "human judgment over raw output." By addressing the underlying cognitive anxiety associated with AI, organizations can foster a more conducive environment for its adoption.
Creating a "Human-in-the-Loop" Framework
A key strategy for mitigating botsitting is the implementation of a "clear human-in-the-loop blueprint." This framework integrates human oversight directly into work processes and decision-making, transforming AI integration from a potential source of stress into a "healthy partnership." In this model, employees transition from being "stressed-out machine caretakers" to "critical editors," actively contributing their expertise to refine and validate AI outputs.
This message needs to be reinforced through comprehensive manager training. Leaders should be equipped to identify red flags, such as the "audit tax" where employees are working longer hours to double-check AI-assisted tasks. They should also be trained to recognize subtler signs of strain, like employees withdrawing from strategic discussions or exhibiting an over-reliance on AI for decision-making, which can indicate underlying issues with AI integration.
Measuring True Value: Beyond Productivity Metrics
To truly assess the success of AI rollouts, Spatz suggests incorporating botsitting considerations into the evaluation process. This involves measuring behavioral, qualitative, and emotional metrics, rather than relying solely on traditional productivity figures. "If an employee saves five hours a week using AI but fills it with five more hours of automated administrative noise, the rollout has failed," she states. "True productivity eliminates friction and frees cognitive bandwidth for work that matters most." This nuanced approach ensures that AI investments deliver genuine value by enhancing, rather than merely shifting, the burden of work.
The challenge of botsitting is not merely a temporary hurdle in the adoption of AI; it represents a fundamental aspect of how organizations can achieve sustainable and beneficial AI integration. As AI continues its relentless march into every facet of business operations, HR professionals stand at the forefront, tasked with ensuring that these powerful tools augment human capabilities, rather than inadvertently creating new layers of complex and exhausting labor. The organizations that successfully build this "human infrastructure" will be the ones that truly unlock the transformative potential of artificial intelligence.
