The root cause is a lopsided investment in artificial intelligence (AI). “It requires the right mindset, the right skill set, and the right tool set,” says Dan Diasio, EY global consulting AI leader. “For the last three years, it’s mostly just been a focus on the tool set.” That imbalance has typically produced a crude calculus where better usage means fewer people—which is precisely what employees fear. Capitalizing on their enthusiasm and earning their trust requires an entirely different approach—one focused less on the tools than on training them to collaborate with agents as colleagues, while making it clear their own roles will evolve rather than evaporate.
03-11-26 | BY FAST COMPANY CUSTOM STUDIO
According to a new report, success
requires a clear strategy—and clear communication to the workforce
How to take full advantage
Clarity may be the most powerful lever organizations are failing to pull. Nearly all (92%) beneficiaries of clearly articulated AI strategies report productivity gains—a 30% jump over their counterparts kept in the dark. “If you don’t have a clear stance as an organization on AI, that’s problem statement number one,” says Kim Billeter, EY global people consulting leader. But the message matters as much as the medium. Leaders must do better at framing AI as the key to growth, not just productivity. “The future of work isn’t about humans versus machines,” Billeter adds. “It’s about humans with AI creating value through collaboration.”
This helps explain why upskilling is happening invisibly, at kitchen tables
late at night. Workers aren’t waiting for their employers to catch up. But this
self-reliance carries a retention risk—employees who receive substantial training with AI are more engaged but also become more willing to leave if
their paths forward are blocked. “If you’re not incentivizing people and giving them tools to use, you’re just giving them all these skills and then they’re going out the door,” Billeter warns. “The key talent you want in your organization is very willing to walk.”
The deeper problem is not the proliferation of AI tools itself, but the mindset that has accompanied it. Three years of tool-centric thinking has left many firms unprepared for the agentic era—an era in which humans are responsible for framing problems, setting intent, and judging outcomes, while agents execute. Without a shared mindset and clear intent from leaders, workers are asked to adapt to changing systems without clarity on purpose, standards, or ownership. Workers encounter AI not as a capability being deliberately built for them but as a series of disconnected experiments happening to them.
Diasio puts a name to the risk that emerges when tools race ahead while leadership intent, human judgment, and agent execution remain unaligned. “AI doesn’t come up with novel ideas,” he says. Without human creativity bracketing the process, companies will produce what he calls “statistical sameness”—outputs indistinguishable from every competitor using the same models. The organizations that thrive using agentic AI won’t be those with the best tools. They’ll be the ones investing in the people on either side of them.
REIMAGINING WORKPLACE RELATIONSHIPS
Commissioned by
Created BY
Commissioned by
Created BY
The vast majority of American knowledge workers are eager to work alongside agentic AI—systems capable of completing tasks on their own—but dread that their new silicon-based coworkers will ultimately put them out of a job (or worse, ask the humans to manage them). That’s the paradox exposed by the inaugural “Agentic AI in the Workplace” survey from the Ernst & Young LLP (EY US) Consulting service line. The poll of more than 1,000 employees of blue-chip firms across a half-dozen industries found that while 84% express enthusiasm for mastering agents, more than half (56%) worry that their productivity will come back to haunt them and that agents will make their jobs obsolete (51%).
That excitement wrapped in anxiety has been compounded by a leadership vacuum. Employees report feeling overwhelmed by new tools (64%) amid a lack of training and clear communication from management about strategy. As a result, they’ve been forced to fend for themselves, learning new skills after hours on their own time (85%). The stakes are high: Organizations risk not only squandering the advantages of agentic AI but also deepening disengagement among their most anxious workers while failing to retain their most confident ones, who may take their self-taught proficiency elsewhere.
logo
+
+
“The future of work isn’t about humans
versus machines. It’s about humans with
AI creating value through collaboration.”
THE SELF-RELIANCE RISK
express enthusiasm
for mastering agents
think
agents
will make
their jobs obsolete
51%
84%
—Kim Billeter,
EY global people consulting leader
“[Gen Z doesn’t]
think of managing these [AI] systems as management.
They’re AI-native
about what the future role of a manager is.”
Generational divides complicate things further, although not in the ways one might expect. While the Wall Street Journal reports that Generation Z is typically portrayed as the most vulnerable to AI displacement, they’re actually the most eager to learn (88%) and least constrained in managing agents. “They don’t think of managing these systems as management,” Diasio says. “They’re AI-native about what the future role of a manager is.” For its part, Generation X evinces a surprising level of confidence (93%), seeing agents as just another wave of new technologies to absorb. Millennial managers are the most worried (72%) that they will soon be swamped by an exponentially growing number of direct reports.
—Dan diasio,
EY global consulting ai leader
Employees’ views on AI agents
The views reflected in this article are the views of the author and do not necessarily reflect the views of Ernst & Young LLP or other members of the global EY organization.
AI-native
collaboration
93%
Gen X
72%
Gen Y
worry they will
be swamped by
direct reports
88%
Gen Z
just another
new technology
to learn
ADOPTING AI, BY GENERATION
This collaboration will in turn
require reimagining workplace relationships. The shift isn’t
one from chatbots to agents, but from delegation to partnership. “They need to be viewed as colleagues,” Billeter argues, “not just tools competing for my job.” For managers, this means developing an entirely new muscle: orchestrating hybrid human-agent workflows rather than simply supervising people.
eager to learn
although they feel vulnerable
to displacement
“That’s a muscle no one has right now,” she says, and it shows: More than half of managers (53%) worry that they won’t be good at it, and a greater majority of nonmanagers (63%) are inclined to avoid such roles in the future.
of agentic AI while keeping
anxious workers engaged