Cliche as it sounds, change has been the only constant for the past half-decade: restructurings, return-to-office cycles, layoffs followed by re-hiring, and the steady arrival of generative AI tools that change what a given role looks like from one quarter to the next. The signals across our latest research and our Generative AI at Work findings are consistent. Employees are running out of capacity to absorb all this change, and their confidence in the leaders steering it is thin.
Change Disruption sits at 55.4% favorable in our latest 3,901-employee panel, the lowest-scoring category in the entire study. Confidence in senior leadership to navigate change reaches only 53.1%. Confidence in leadership's ability to create an adaptive culture sits at 52.3%. Roughly half the workforce is reporting, in different ways, that leadership does not seem equipped for what is coming.
Generative AI offers the most obvious current test of that readiness. The anxiety it creates differs from the familiar "we are restructuring" anxiety, which has a clear endpoint and a defined cause. GenAI is the kind of uncertainty where the answers are not fully known yet, the pace is accelerating, and the people running the organization are working it out alongside everyone else.
Most change management approaches assume change is fundamentally a communication and strategy problem, to be solved with clearer messaging, more town halls, and a stronger overall business case for transformation. Those tools work where the change is well-defined and the destination is known, and they lose traction where both are shifting weekly.
The item "my workplace avoids stagnation" sits at 67.0% favorable across the full sample, meaning roughly one in three employees already feels stuck before any new change initiative arrives. That is the baseline state from which most organizations are now trying to roll out GenAI strategies, new operating models, and accelerated workforce planning.
What traditional change frameworks miss is the moment-to-moment human behavior that determines whether people meet uncertainty with curiosity or with fear. That behavior is what improvisational theater has spent a century training, and what the new Perceptyx and Second City Works research discuss here measures at scale.
The 32.1% of employees who experience all five improv principles report substantially different perceptions of their leadership and their workplace under change. The principles appear to reshape how employees interpret leadership behavior under uncertainty, not only how they feel about the work itself.
| Measure | Improv group | Remainder | Gap |
| Change Disruption (category) | 84.0% | 40.1% | +43.9 pts |
| Confidence: senior leadership can navigate change | 82.0% | 37.3% | +45.3 pts |
| Confidence: leadership can create adaptive culture | 81.7% | 36.5% | +45.2 pts |
| Confidence: direct manager adapts to change | 85.7% | 44.0% | +41.7 pts |
| Change & Innovation benchmark | 94.4% | 50.9% | +43.5 pts |
Employees in the improv-principle group are more than twice as likely to express confidence in senior leadership's ability to navigate change (82.6% vs. 37.3%). The same leadership, in the same conditions, is interpreted as substantially more competent by employees who work in improv-informed teams. The mechanism is interpretive rather than emotional: how employees read leadership behavior under uncertainty shifts measurably when the behavioral grammar of the team is improv-informed.
Humor as a resilience tool (the improv principle “Laugh”). The item "humor builds resilience during challenges" reaches 71.4% favorable overall and 100% among the improv-principle group. Humor in this context does not function as a way to dodge hard conversations. It works as a mechanism for processing difficulty without shutting down, which keeps cognitive flexibility intact when circumstances become threatening. Teams that practice it adjust to disruption faster than teams that don't.
"Yes, and" reshapes how employees interpret leadership (the improv principles “Improvise” + “Ensemble”). Future Vision scores reach 95.8% favorable in the improv-principle group against 58.3% for the remainder, a 37.5-point gap. The item "senior leadership communicates a clear vision" shows a 41.9-point gap (95.0% vs. 53.1%). In environments where building on others' ideas is the operating default, top-down decisions land differently. Employees who have internalized a generative orientation toward new information find logic and intent in leadership decisions even when those decisions are imperfect. The same email from the same CEO reads as direction in one team and as confusion in another, and the behavioral grammar of the team is what determines which.
Ensemble as the antidote to isolation during change (the improv principle “Ensemble”). Well-Being category scores reach 94.5% in the improv-principle group vs. 62.3% for the remainder (+32.2 pts), and the item "organization cares about my health and well-being" reaches 93.4% vs. 54.9% (+38.5 pts). What employees are responding to here goes deeper than any wellness program, landing instead at the relational level of being seen and supported, which improv principles appear to reliably create. Change is destabilizing in part because it isolates people, and ensemble behaviors keep the team connected when the work around them is in flux.
The Change Disruption category sits at 55.4% favorable in our data. Regrettably, as more roles get reshaped by AI tools and more employees see colleagues using AI in ways that change what their own work looks like, that 55.4% has nowhere to go but down unless something shifts in how organizations support people through it.
The improv-principle group has already cleared the hurdle. On questions about competitive readiness, the gap is striking: 96.6% favorable in the improv group vs. 63.5% for the remainder on the item "the organization is doing what is necessary to compete effectively" (+33.1 pts). Employees who work in improv-informed teams interpret their organization as more capable of meeting the moment, which is itself a workforce capability that compounds over time.
GenAI is the vehicle for this conversation rather than the subject of it. The focus must stay on human behavior, because the durable capacity to keep working productively under ambiguity is what an AI-saturated workplace requires from its people, and most workplaces have not yet built that capacity.
Two-thirds of the workforce is measurably less equipped to navigate change, because they work without the five improv principles in place. The behaviors that close that gap are observable, learnable, and reinforceable, which makes them addressable through deliberate practice rather than left to luck or hiring filters.
The resilience effect originates with daily habits at the top, not with HR program design alone. Improv principles operate at the level of how leaders run meetings, how managers respond to risk, how teams handle bad news, and how decisions get communicated in conditions of incomplete information. A bolt-on training module delivered around the side of an existing change initiative will not produce the effect the data describes. The leaders who model the behavior themselves are the ones whose organizations get the resilience benefit.
For HR, the starting point is the same as in our companion piece on engagement: measure what is actually happening at the behavior level. The behavioral signal will not appear in averaged engagement scores. It surfaces in the items that ask about humor, building on ideas, risk tolerance, and felt support. Our Generative AI at Work research details how those signals connect to the specific anxieties GenAI is now driving across the workforce.
Perceptyx provides the listening infrastructure to establish a baseline, connect employee experience signals to business outcomes, and track behavior change over time. The Second City Works provides the experience design, facilitation, and habit-formation methodology that builds these behaviors at the team and manager level.
A full report, with findings across all five improv principles and every dimension of employee experience, is coming later this year. To read the first part of our research on improv principles and employee engagement, click here.
To see what the data show about employee experience in the GenAI era, download Perceptyx's Generative AI at Work report or schedule a 30-minute demo with our team.