Read a handful of tech headlines, or even spend a few minutes on LinkedIn, and you’ll likely see the same hurdle to AI adoption cited over and over: a lack of employee buy-in, especially among workers who fear AI agents might one day be able to do large portions of their current jobs.

But pushback from employees is not holding back agentic AI, according to a survey of The Information readers. In fact, readers say, workforce resistance ranks at the very bottom of the list of barriers to adopting the technology within their organizations. Rather, agentic AI efforts are currently limited by a mix of technical, security and organizational challenges, and many readers say their organizations have struggled to identify near-term use cases that yield positive value.

Despite these hurdles, readers largely see AI agents eventually increasing operational efficiency, reducing costs, and providing other business benefits. While respondents dismiss employee resistance as a significant hurdle, the survey of 281 readers did reveal significant challenges in reskilling and upskilling employees to work alongside AI agents, with readers citing the pace of change, time constraints and a lack of training resources. This may signal an opportunity for companies to double down on their AI education efforts.

Leaders should also consider how organizational, technical and talent-related challenges intersect. For instance, the immature nature of current offerings may threaten to lessen employee enthusiasm for AI agents.

As one reader puts it: “Lack of success with adoption leads to [a] lack of enthusiasm to learn.”

Agentic AI: Key Uses and Benefits

Overwhelmingly, readers expect that AI agents will increase operational efficiency in their organizations, and half or more cite cost reduction, innovation and enhanced decision-making among the most anticipated benefits.

Jake Burns, AWS executive in residence, says that efforts aimed at boosting innovation may have the highest long-term potential impact. “Using AI can certainly help you create incremental cost savings, but you’ll get far more value by using it to create new capabilities—capabilities that were previously considered impossible,” Burns says.

Asked to identify specific tasks or departments that they envision AI agents supporting, readers cite a wide range. Task automation and workflow automation takes the top spot, followed by data analysis and insight generation. Around one-third of readers envision AI agents supporting the following tasks: service and support; employee assistance and productivity enhancement; content creation; and application development.

Burns calls data analysis a “no-brainer” application of AI. “One of the superpowers of this technology is that it can be used to analyze vast amounts of data and identify insights in real time,” he says. “Every organization in the world should be utilizing that.”

For organizations that have struggled to find value in AI, Burns says, the problem is often that leaders and employees treat the technology like an “easy button,” rather than investing in training. “If you treat it like an easy button, you’re going to get bad results, and you’re going to conclude that it’s useless,” he says. “I believe this explains much of the negative sentiment. Its usefulness is proportional to the skill and effort that you put into it, and those who use it well are getting incredible value. But it’s important to recognize that this does require effort.”

Adoption Barriers & Challenges

Starting at the bottom: Only 14% of readers say workforce resistance is a primary barrier to adopting AI agents in their organization.

Although this may surprise some, the finding mirrors other research showing that many workers are at least somewhat open to AI agents. A Stanford study shows that only 23% of workers fear job loss due to AI automation, with 69% saying they welcome automation that would free up their time for higher-value work.

Burns notes that many employees have enthusiastically adopted AI tools in their own lives—using ChatGPT instead of Google to look up information, for example. However, he adds, workers often bristle at company mandates to use the technology for work, especially when companies launch poorly designed AI pilots without proper training or context. “We should be cautious not to interpret resistance toward specific implementations as resistance toward the technology itself,” Burns says. “When implemented well, AI removes pain points from employees and amplifies their current skills.”

Readers say the top barrier to adopting AI agents is a concern about data privacy and security, followed closely by a lack of validated agentic AI products in the market and an inability to identify ROI-positive use cases.

Here again, some readers express skepticism about the utility of AI tools. In an open response answer, one cites the “unreliability of LLMs beyond POC stage,” with another citing “prevalence of hallucinations/mistakes.” However, survey data indicates that most leaders see value in agentic AI; only 15% say “a lack of belief or trust in the technology among company leadership: is a top adoption barrier.

“There is certainly a divide among the executive teams I engage with,” Burns says. “Many don’t recognize how quickly the technology is advancing, but the ones that do are able to clearly see how much of a competitive advantage it is now giving them.”

Integrating AI Agents Into the Workforce

Asked about their biggest concerns surrounding the integration of AI agents into their organization’s workforce, readers again largely dismiss the impact of employee resistance. Only 20% of readers cite cultural resistance to AI adoption among their most significant concerns, and even fewer (19%) say they are concerned about potential job displacement. Once again, these factors rank at the very bottom of the list of options.

When it comes to integrating AI agents into the workforce, readers are much more concerned about technical limitations. The vast majority cite the reliability and accuracy of AI outputs as a top concern, and a majority cite data privacy and new cybersecurity vulnerabilities. “Data privacy, security, reliability and accuracy can be major challenges,” Burns says. “That’s why it’s so important to run AI models in an enterprise-grade environment with the right guardrails, security controls and observability tools.”

Concerns about the impact on people also show up, with a majority of readers citing concerns about overreliance on AI decision-making and more than one-third of readers citing challenges in AI-human collaboration and a loss of human skills or expertise.

Preparing fro AI Agents & Upskilling Workers

Organizations are taking a number of steps to prepare their culture for the introduction of AI agents, including launching pilot projects to demonstrate the benefits of AI, identifying internal AI champions and involving employees in AI implementation.

Still, readers foresee a number of challenges in reskilling and upskilling their organization’s workforce to work alongside AI agents. The most common are the pace of change/constantly evolving AI offerings, time constraints and a lack of suitable training resources or partners.

Burns says organizations have the right idea relying on internal champions to help train their colleagues. To identify these champions, he says, leaders should look for employees who have not only shown enthusiasm for AI, but those who have demonstrated real results. By providing employees an enterprise-grade AI “sandbox,” he says, organizations can create the conditions needed for internal champions—and valuable business cases—to emerge.

“You don’t always know who these people are going to be,” Burns says. “You want outside-the-box thinkers, rebels and those unafraid to experiment. You’ll find out who they are by giving everyone the opportunity and then by seeing who takes advantage of it. You don’t choose them. You discover them.”

Conclusion: Seizing the Agentic AI Moment

The Information’s survey is clear about the factors limiting agentic AI breakthroughs within organizations. For most, employee resistance is not a significant challenge. Rather, companies are struggling to overcome organizational uncertainty, technical complexity and product immaturity at a time when it seems like technology capabilities and business opportunities change by the week.

Among the survey’s most important findings:

• Employee resistance is not the problem. Only 14% cite employee resistance as a top barrier, and fewer than one-in-five respondents are concerned about potential job displacement. This suggests that many employees will likely embrace agentic AI solutions that improve their productivity and free up time for higher-value work. 
• Technical and organizational challenges dominate. Employee acceptance alone doesn’t guarantee agentic AI success. Today, adoption is limited by security concerns, product maturity and unclear use cases. Organizations need enterprise-grade AI platforms to minimize risk.  
• Workforce enablement is urgent. To take advantage of employees’ current willingness to adopt agentic AI, organizations must conduct skills assessments, launch training programs and create hands-on opportunities for experimentation and implementation. 
• The window is open. Organizations are already taking action: investing in pilots, identifying internal champions and implementing governance frameworks to lay the foundation for agentic AI success. While challenges remain, companies that arm their employees with effective AI agents and training are poised to increase efficiency, cut costs and spark innovation. 

Methodology

This report is based on a survey of 281 readers of The Information, conducted in August 2025. Respondents represented a wide range of industries, company sizes, functional areas and titles.

Industry: By far, the largest industry represented in the survey was technology, media and communications, which comprised 43% of respondents. Seventeen percent of respondents work for professional services companies, 13% work in financial services/capital markets and 6% work in healthcare and life sciences. The remaining respondents represented private equity, hospitality and leisure, industrial manufacturing, retail and others.

Company Size: Fifty percent of respondents came from companies with annual revenues under $10 million; 21% had revenues of more than $1 billion; 17% had revenues between $10 million and $100 million; 8% had revenues between $100 million and $500 million; and 5% had revenues between $500 million and $1 billion.

Functional Area: Respondents represented multiple functional areas, with the largest groups coming from general management (23%), information technology (15%), engineering (14%) and marketing and communications (13%). Other areas, each comprising under 10% of respondents, included research and development, sales, finance and legal.

Title: The most represented job title among respondents was CEO/Owner (30%), followed by director (20%), employee (15%), SVP/VP (11%) and manager (8%).