Artificial intelligence agents are supposed to augment workers, relieving them of low-level mundane tasks and boosting their insights for performing higher-level tasks. Most employees recognize the potential of agents, but when it comes to learning and adopting them, they’re left on their own. At the same time, many are avoiding potential management roles, because no one really knows yet how to manage human-AI hybrid workforces.
That’s the word from EY, which just released a survey of 1,100 professionals and white-collar workers that finds nervous enthusiasm and confusion about the advent of AI agents in their workplaces.
The stakes are high. “Our agentic AI future is nothing if not uncertain, with trillions of dollars’ worth of bets on the line from the biggest players in technology,” the survey’s co-authors, led by Dan Diasio, EY Global Consulting AI Leader, stated.
In a word, for leaders and employees alike, it’s all simply overwhelming. Close to two-thirds of workers who use agentic AI, 64%, reported they are overwhelmed by the amount of new agentic AI tools being introduced at their workplaces.
Tellingly, the rise of AI agents is creating great uncertainty as to what it means to be a manager as well. Until recently, becoming a manager meant being a team leader and organizer of people. A majority of non-management employees in the study, 63%, are reluctant to aspire to management roles which require overseeing AI agents in conjunction with human workers.
Going forward into hybrid workplaces, managers need to address some pressing questions:
- “How does it work when your team is half humans and half agents?
- “How will your people feel about collaborating with AI — will they be energized or hostile?
- “Will you need to give an agent feedback?”
Even current managers and supervisors do not feel confident taking on these questions, “even though 88% are convinced that AI agents will always require human managers’ oversight,” the EY team asserts.
Still, there’s agreement that agentic AI will have a positive impact on productivity, efficiency, and daily work experience, the survey shows. But many are left to experiment with the technology on their own. Most, 85%, say they are learning about how to work alongside AI agents outside of work. 83% state that most of what they know about working with agentic AI is self-taught.
When it comes to leveraging AI agents, too often workers’ curiosity “is met with silence: little training, limited communication and no clear roadmap,” the co-authors explained. “This is where enthusiasm risks curdling into resistance.”
About 52% of senior leaders insist their organizations have “a fully deployed
initiative in investing in agentic AI training or upskilling” for their employees. Apparently, there is still a lot of work to be done in this area. But leaders seem to be in the dark about how well their AI agent strategy, if any, is percolating throughout their organizations. At least 23% of employees below the VP level state their manager or supervisor has not clearly communicated the organization’s strategy on AI agents, compared with just 9% of those at the VP level or above.
“While some jobs may be more immediately impacted, if we prepare ourselves to use AI as a powerful tool, we can build the skills to do more meaningful work – and even create roles that don’t exist yet,” said Diasio. “Agentic AI should be seen as a thought partner that sharpens human judgment and expands capability – not as an easy button that risks trading brainpower for short-term productivity.”
Diasio and his co-authors offer suggestions to boost confidence in AI agent deployments:
Push back on the narrative of doom-and-gloom on job security and commit instead to market-defining growth: Leaders and managers should talk about ‘what can be achieved through new products and services, new business models and more forms of talent working together enabled by AI, not replaced by it.”
Be open about AI plans and vision: “The survey results highlight a clear
gap between talking about AI and doing AI,” the co-authors state. “And some organizations aren’t doing either very well.” Significantly, where there is reportedly clear communication on AI strategies, 92% of employees report that working with AI agents has positively impacted their team’s productivity – “a rate that is 30 percentage points higher than their counterparts at companies without clear communication.”
Step up training: Lack of appropriate training has vexed employees since the advent of technology, and the current AI era is no exception. The challenge this time around is training people for technology that changes on an almost weekly basis. Many companies don’t even bother, the report states. Instead, as the survey shows, workers are relying heavily on self‑directed learning on their own time as well as their peers and managers. The trouble is that “untrustworthy sources have rushed in with poor lessons that are potentially more risky to your people than useful.”
With AI in all its forms, we’ve all been on a journey of discovery, with both the hopes and fears that accompany it. This survey helps illustrate the crucial role tried-and-true management continues to play.