Work, at first glance, seems easy to define. It is what people do in exchange for pay. Yet, like many concepts that appear simple, it dissolves under deeper scrutiny. Consider the many individuals who are visibly at work but not obviously working: the manager perpetually in meetings that produce no decisions, the employee toggling between tabs with Olympic dexterity, or the executive whose calendar is full but whose contributions are hard to trace. A recent survey found that 65% of employees admit (the actual figure is probably higher) engaging in regular “productivity theatre”, as in performing tasks to appear busy without actually doing meaningful work. Conversely, there are those who do substantial work without formal recognition or pay. Economists have long noted that unpaid labor, particularly domestic and care work disproportionately performed by women, would amount to over 10-trillion dollars annually if accounted for in GDP (that’s around 9% of GDP).
The ambiguity does not end there. Even when we agree on who is working, we struggle to measure how much. Time, the most common proxy for effort, is a notoriously poor indicator of productivity. Hours logged are weakly correlated with value created. The whole point on hiring for talent is that it allows some individuals to achieve more with less, or produce higher levels of value and output the same effort or less input. One employee may produce in two hours what another cannot deliver in ten. Meanwhile, the modern knowledge economy has made output itself harder to define. Unlike factory work, where units produced can be counted, much of today’s work involves intangible contributions such as ideas, judgments, or influence.
Faced with this measurement problem, organizations oscillate between two flawed approaches. At one extreme, they track granular, observable behaviors: keystrokes, emails sent, calls made, minutes online. These metrics have the virtue of being measurable, but they prioritize quantity over quality and invite gaming. At the other extreme, firms assess broad outcomes such as revenue growth or profitability. Yet these outcomes are often detached from individual contributions. Claiming “we increased profits during my tenure” says little about causality or personal impact. Between these poles lies a persistent gap between what is measured and what actually matters.
As if things weren’t complex enough, we can add politics to the mix. Work has never been purely about output. It has always involved signaling, perception, performance, and reputation management. Erving Goffman, the father of modern sociology, famously described social life as a form of theater, where individuals perform roles to shape how others perceive them. The workplace is no exception. Employees engage in what Goffman would call “impression management,” curating an image of productivity and competence by putting on a highly choreographed act, making an effort to “fake good”.
Sometimes this performance becomes internalized. In what he called “deep acting,” individuals convince themselves that they are fulfilling their role effectively, even when the objective impact is unclear. This helps explain a well-established finding in organizational research: self-rated performance correlates only modestly with others’ evaluations. Many people genuinely believe they are high performers, even when evidence suggests otherwise.Any manager who has had to deliver a difficult performance review will recognize this gap firsthand: the need to provide a reality check to employees whose self-perception is far more favorable than their actual impact.
There is also the question of meaning. The late anthropologist David Graeber argued that a significant portion of modern employment consists of “BS jobs,” roles that even the people performing them suspect are pointless. In such cases, one can be both productive in a narrow sense and disengaged in a broader, existential one. The paradox is that work can be simultaneously effortful and meaningless.
The Performance Paradox
When individuals spend more time focusing on the performative aspects of work than on their actual job performance, their job performance declines but their career advancement benefits. In such environments, being seen to perform becomes more valuable than value itself.
This raises an uncomfortable possibility: those who invest more time in appearing productive than in actually producing value may outperform those who do the opposite. Consider two employees. One focuses relentlessly on delivering high-quality work but neglects visibility. The other spends considerable effort ensuring their contributions are seen, narrated, and aligned with organizational priorities, even if the underlying output is less impressive. In many contexts, the latter will advance faster. Organizations reward not just value creation, but value perception – perceptions trump reality. Careers are built as much on signaling as on substance.
The logic is not entirely irrational. Managers operate under constraints of limited information. They rely on proxies, narratives, and signals to infer performance. In this environment, impression management is not merely deception. It is adaptation. The problem arises when the system systematically rewards the appearance of productivity over its reality.
Complicating matters further, managers themselves are often subject to the very same dynamics. Many are equally preoccupied with signaling performance upward, curating their own narratives for senior leaders rather than objectively scrutinizing their teams. When evaluation becomes a game of managing impressions at every level, the entire system drifts away from substance and toward theater.
Will AI Change Anything?
Enter artificial intelligence, which complicates matters further. Generative AI has made it possible to achieve the same outputs in a fraction of the time. A report that once took a day can now be produced in an hour. A presentation that required a team can be assembled by one individual. In principle, this should lead to a productivity boom. In practice, the effects are not as visible.
As the World Economic Forum reported, in many organizations, adoption rates of generative AI tools now exceed 60 to 70 percent of employees. Yet aggregate productivity gains remain modest at best. The reason is not that the technology fails, but that its benefits are often privately captured and poorly redeployed. Individuals use AI to complete the same tasks faster, but instead of reinvesting the time saved into higher-value activities, they simply do the same work with less effort. The result is a form of concealed efficiency rather than visible output.
At the same time, firms are experimenting with “work without jobs,” replacing fixed roles with fluid, skills-based assignments where people move across projects rather than sit in defined positions. In theory, this sounds adaptive and modern. In practice, AI may be revealing the opposite problem: not jobs without work, but jobs that contain surprisingly little real work to begin with.
As a result, AI is often used clandestinely. Employees quietly leverage these tools to reduce effort without necessarily increasing output. From their perspective, this is rational. If performance is judged by visible deliverables rather than underlying effort, there is little incentive to disclose efficiency gains. The result is a peculiar form of “hidden productivity,” where individuals do the same work faster but do not expand their contributions accordingly. This is productivity without value creation, or at least without value realization.
For managers, this creates a dilemma. Should they reward outcomes regardless of how they are achieved, or should they attempt to monitor process? The former risks encouraging minimal effort and underutilization of human potential. The latter risks reverting to intrusive surveillance and low-value metrics. The more promising path lies elsewhere: redefining expectations. If AI enables employees to do more in less time, then performance standards should evolve accordingly. The goal is not to preserve effort, but to expand impact.
To be sure, performative work is not the exclusive domain of employees. Leaders are equally susceptible. Executives engage in their own forms of impression management, signaling strategic vision, decisiveness, or busyness. Town halls, offsites, and elaborate presentations can become rituals that convey activity without necessarily driving outcomes. Just as employees may pretend to work to impress their bosses, bosses may pretend to lead to impress their stakeholders, or even each other.
The symmetry is telling. Organizations are not just systems of production. They are systems of perception. At every level, individuals are incentivized to manage impressions as much as to create value. The risk is that the entire enterprise drifts toward theater, with diminishing connection to substantive outcomes.
What can be done?
What, then, is to be done? The central challenge for organizations is to reduce the gap between individual career success and actual value added. This is easier said than done, but several principles can help.
First, focus on outputs that are as close as possible to value creation, even if imperfect. Resist the temptation to rely on easily measurable but low-value proxies. Second, triangulate performance using multiple sources of data, including peer and customer feedback, to counteract individual biases and impression management. Third, create transparency around how work contributes to broader outcomes, making it harder to claim credit without substance. Fourth, update performance expectations in light of AI-driven efficiency gains, rewarding those who reinvest saved time into higher-value activities rather than those who simply maintain the status quo.
Finally, and perhaps most importantly, acknowledge that some degree of performance theatre is inevitable. The goal is not to eliminate it, which would be unrealistic, but to align it more closely with reality. In other words, to ensure that the art of appearing productive does not overshadow the substance of being so.
Work, it turns out, is not just about what we do. It is also about what we show, what we signal, and what others believe we have done. In the age of AI, this distinction becomes even more consequential. The organizations that thrive will not be those that eliminate pretense altogether, but those that ensure it serves, rather than substitutes for, genuine value creation.