Scaling AI has defined many organizations’ goals this year, and the push is visible in the data.
McKinsey’s 2026 State of AI survey found that 44% now report that AI is scaling across their enterprise, up from 38% a year earlier. Financial results have stayed flat over the same period: the share of respondents reporting that AI has contributed to their organizations’ EBIT is essentially unchanged from a year ago. A growing body of research suggests that much of this gap comes down to the people expected to use the tools.
PwC’s recent survey found that while AI use rose 10 points in a year, only half say they can get the training they need, down from 59%. Gallup’s tracking points to a leadership gap as well.
As of May of this year, 25% of U.S. employees say their organization has communicated a clear plan for integrating AI into current practices, and only 36% in AI-integrating organizations strongly agree that their manager supports their team’s use of AI.
Workers have noticed.
A Jobs for the Future survey released in March found that 56% of employees report their employers have not consulted them on AI implementation, and that just 36% of workers now say they have the training and resources they need to use AI in their jobs, down sharply from 45% a year ago.
Shinichiro (SHIN) Nakamura, President of one to ONE Holdings, sees these numbers as a leadership challenge at least as much as a technical one. In an interview with the executive, he explained where AI training tends to fall short and what leaders can do to earn employees’ trust and prepare them for the work ahead.
Most organizations say scaling AI is a top priority. Why isn’t that translating into progress?
Simply put, priority doesn’t necessarily mean progress. One report found that only a third of workers have received employer-provided AI training in the past six months. At the same time, there’s growing concern from workers that the more involved AI becomes at work, the more likely their roles could be replaced, and there’s little clarity on what work will look like with more automation.
That uncertainty breeds mistrust, which is a serious barrier to scaling AI. Without trust, transparency, clarity, and collaboration between leaders and teams on the ground, organizations will struggle to build out their AI strategies.
When AI adoption stalls, leaders often blame worker resistance. Is that fair?
Resistance exists, but it’s only a small part of the wider problem. In most cases, what looks like resistance is actually a leadership failure to explain why and how AI tools benefit the people using them.
Individual adoption has actually boomed in the past year, and Gallup reports that 52% of workers are using AI in their role. So persistent training and adoption failures can’t be entirely pinned on negative attitudes among workers. The crux of the issue is how these tools are being integrated into workflows and introduced to teams.
Younger workers, particularly Gen Z, seem to adopt AI faster. Doesn’t that suggest older workers are more resistant?
Not necessarily. I think the generational gap points to a deeper organizational issue: interfaces and training aren’t being tailored to workers’ varying needs. It’s not just about what specialty or department you work in, but also how long you’ve been working, the kind of experience you bring to a position, the key learnings you’ve had throughout your professional life.
Age, experience, and exposure are important factors to consider beyond technical focus areas when developing AI training programs. UK-based research shows that the adoption gap across generations is a symptom of inadequate training, not age.
If leadership’s approach is the real barrier, what are organizations getting wrong in their training programs?
I see three major mistakes more often than not. The first is another misdiagnosis: early AI performance problems get dismissed as a people problem, as if workers simply can’t get to grips with the tools. The real culprit is usually a lack of groundwork in data management, orchestration, and interoperability. New systems underperform because of poor data or opaque usage history, which piles extra work onto employees who have to correct outputs and keep feeding in context. When tools become a burden, that invisible labor quickly drains enthusiasm for them, and training that doesn’t account for this leaves employees feeling blindsided.
The second mistake is a lack of coordination, specifically, no shared glossary. Without standardized terminology that departments and individuals can rely on, outputs on a busy factory floor or a fast-paced office become unpredictable. Workers start to associate AI systems with unreliability, even though the big selling point of these tools has always been strong results, delivered faster. That shared terminology should be taught in initial training, but it also needs to be always accessible and kept up to date. Controlled vocabulary around AI use cases and systems should be treated as an ongoing maintenance task, owned by HR and department heads, in addition to the IT team.
The third error is leaning too heavily on theory; it isolates the tools from real working conditions and doesn’t prepare people adequately for overseeing AI or intervening when something goes wrong. A concept can be neat and make total sense on paper, but a lot messier in real life. There are complications or scenarios that are impossible to prepare for with theory-based learning alone. Training should sit as close to the actual workflow as possible. I’m a firm believer that practical exposure is the best way to familiarize employees with AI, according to how the organization wants (and needs) them to use it.
Beyond training, what does it take to build trust and sustain adoption?
Strong communication. It’s the thread connecting boardroom expectations and workers’ realities, and it helps AI deployment scale securely and productively. It matters most in busy operations with overwhelming production lines involving multiple vendors, teams, suppliers, and departments.
Ask the right questions of the right people before implementing a tool. When you ask employees about their pain points and workflow constraints ahead of time, the change feels like a response to their input rather than something imposed from above. That gives them a sense of ownership and involvement, which fosters trust and transparency. It echoes Alfred Adler’s idea of respect: when people feel encouraged, capable, and appreciated, they act cooperatively and collaboratively. That’s a vital foundation for any change in communication or digital transformation.
How should firms prepare their workforce for an AI-powered future?
They need to build an operating model for a workforce that will keep adapting as AI’s role expands, not treat AI as a one-off implementation. The first step is matching ownership with the tool. Process-specific AI is very different from general-purpose AI, so it requires structured input from the employees doing the work before any change hits operational workflows. It’s a simple approach, but one that’s often overlooked in the rush to deploy. In manufacturing, for example, operators, engineers, and quality control each bring a comprehensive view of frontline realities.
Change breaks when decisions rest on assumptions instead of documented details and experiences from the people doing the work. Written materials and recordings are useful for building consensus and gathering feedback ahead of time, whereas disorganized verbal feedback lets important information slip through the cracks. Documentation also creates an auditable practice that HR, managers, and other leaders can track.
Build a dedicated support team to prepare detailed workflow documentation before engaging other employees. It softens disruption and lays the groundwork for smoother integration, with minimal abrupt impact on people’s roles and work.
Finally, what cultural shift matters most?
Repositioning the role of leaders. They shouldn’t only spearhead strategy, but also be dedicated to supporting, preparing, and extending employee capabilities with AI. A top-down mandate won’t nurture trust or suppress resistance. Collaborative, empowering leadership is what sustains enthusiastic adoption over time. Ultimately, organizations must treat AI as more than another digital transformation project. It’s a strategy that requires careful change management, talent management, clear communication, and strong leadership involvement.
