Employees are spending 6.4 hours a week micro-managing outputs, fixing errors, and repeatedly feeding in context to AI tools, according to a Glean Work AI Institute report. That’s almost a full day. People are left asking: What about my real expertise and job responsibilities?
Frustration is quietly building and becoming a retention risk. Glean also found that those who spend a large share of their AI time on botsitting tasks are 73% more likely to be actively looking for another job.
It’s easy to pin this on AI limitations, but there’s a deeper underlying challenge: a workforce skill gap most organizations didn’t see coming, and haven’t prepared for.
What Is Botsitting?
Botsitting is the act of making AI output usable. It includes feeding AI tools context, checking outputs, and revising and cleaning up mistakes. Glean Work AI Institute coined the term in its report, but it caught on quickly because it names a previously unnamed phenomenon employees are experiencing every day.
Botsitting at Work and the Productivity Conundrum
Most companies now require (or at least encourage) employees to use AI at work. But the outputs often disappoint, and the cost in time and effort is growing. As an example, almost two-thirds (63%) of C-Suite and VP-level leaders report redoing work that was too AI-reliant, according to Founder Reports, and nearly half of all employees have fixed or redone a colleague’s for the same reason. The Glean study found that, “for every hour a worker spends getting useful output from AI, they spend roughly another hour making it usable.” That’s a productivity tax that wasn’t in any AI rollout plan.
The rise of this tedious review work leaves employees feeling like they are babysitting a novice colleague whose work they are responsible for, instead of doing the work they were hired to do. To add to that, only 13% of employees say their organization is performing significantly better because of AI use.
All that effort is producing minimal business outcomes to date. The quality of AI output is only part of the problem. The real issue is older and more familiar than most organizations want to admit: they deployed new technology without preparing the people who have to use it.
What employees are feeling is the chafe of the skill gap and the discomfort of the early stages of learning. No one prepared the workforce for how to work with AI tools well, and most can’t learn fast enough.
Now, it’s the responsibility of the business to fill the gap.
The Skills Employees Need to Collaborate with AI
Most organizations prepared their workforce for AI adoption by explaining what the tools could do, and maybe even how to use them. But knowing a tool’s capabilities doesn’t mean employees have built the skills to leverage those capabilities effectively.
Working alongside AI requires a different skill set than most job descriptions cover. While every job role and industry is unique, three universal areas stand out:
Technical auditing. When an AI model produces a mistake, employees need to be able to spot it. When a prompt produces a poor result, they need to know how to improve it. These are learnable skills. Most organizations just haven’t built structured pathways to develop them. This is especially true for individual contributors who may not be as familiar with reviewing work for quality and accuracy.
Creative judgment. AI generates outputs quickly. The real skill is knowing which outputs are worth keeping and which need to go back to the drawing board. This requires domain expertise and the ability to evaluate work in context. AI models are not inherently creative or original. They pull from existing patterns and norms, which means reviews need to check for more than surface-level accuracy. They need a human hand.
AI prompting and system fluency. Knowing how to communicate with an AI tool, how to give it the right context, and how to refine its behavior over time is itself a competency. It requires knowing how AI systems work. Most employees are developing this capability through trial and error, which is exactly why the work feels tedious and unrewarding. It also takes time to get right, despite vendors often presenting AI tools as out-of-the-box solutions.
If organizations skip building these skills, employees end up compensating for a preparation gap they were never told they had. They feel like they are babysitting a robot, rather than exercising new skills like creative judgment, technical auditing, and AI prompting. Why? Because the early stages of learning new skills are almost always uncomfortable, and without a clear path forward, that discomfort can evolve into frustration. But businesses can make it easier.
How L&D and HR Can Promote AI Fluency to Beat Botsitting Fatigue
Employees who feel like botsitters are also a retention risk hiding in plain sight, as most are already more likely to be looking for another job.
The solution is to build AI fluency across the workforce, and it’s going to require those who hold the people data, the skill data, and the budget (HR, L&D, and the C-suite) to align and move together. AI tools work best when employees know how to use them, and train them over time to work even better.
Here’s where to focus:
Personalize development. According to Deloitte, businesses taking a personalized approach are almost three times as likely to report better business and human outcomes for change initiatives. For example, general AI training doesn’t close the gap for role-specific skills. The skills an HR analyst needs to work with AI differ from what a software engineer or a sales manager needs. People can learn faster when all their learning content and communications are tailored to them. In fact, that same Deloitte study found that employees rank personalization and customization as useful when trying to adapt to change.
Effective AI fluency programs map role-specific skill gaps first, then deliver targeted development against those gaps. Organizations like TEKsystems and ZS go further, using AI coaching and roleplays to personalize development at the individual level.
Connect the technology ecosystem for better skill intelligence. A significant portion of botsitting time goes toward moving information between tools that don’t communicate with each other. Employees become the bridge between disconnected systems, often manually moving context into an AI platform. Simplifying the tech stack reduces the manual bridging work. When learning is integrated into the flow of work, AI tools get the context and skill data they need to actually perform, and development tailors itself to the individual.
Measure change with real business outcomes. Once organizations connect and surface skill intelligence, they can connect learning to bigger-picture outcomes. Part of botsitting fatigue comes from watching effort disappear into busywork with no visible business impact. This is solved by connecting development directly to key business goals so everyone can see the results. For example, Capgemini uses real project data as skill data, connecting learning activity directly to performance and giving leaders an accurate view of workforce readiness. Estelle Maione, Global Head of Learning and Capgemini University, says that, even if day-to-day learning is invisible and embedded into the flow of work, the outcomes should always be visible.
The Price of Botsitting AI Platforms
The growing frustration with botsitting is proof of the human-AI capability gap. Technology is moving faster than most workforces can adapt, and most haven’t been prepared for effective AI collaboration.
The result is hours spent managing errors and redoing what AI got wrong. It’s tedious work that no one prepared employees for, requiring a set of skills that most organizations still haven’t built. Closing those skill gaps is urgent, both to keep pace with AI transformation and protect engagement and employee retention. That requires building AI fluency fast alongside structured opportunities for practice, unlocking skill intelligence across the organization, and measuring how that growth shows up (or doesn’t) in business outcomes.
Vamos manter contato.
Quero assinar a newsletter mensal que oferece insights exclusivos, anúncio de eventos e novidades sobre as soluções da Degreed.