Your People Are Not Struggling With the Technology. They Are Struggling With What It Means.
The pattern is familiar. An organisation invests in a new AI tool, rolls out training sessions, creates documentation, and opens a support channel. Adoption starts promisingly. Then it plateaus. Usage drops. Teams revert to old processes. Leadership looks at the numbers and concludes that the training was insufficient, so they schedule more of it.
This cycle repeats across industries, and it almost always misses the point.
According to Gartner's 2025 research on digital workplace adoption, resistance to new technology in the enterprise correlates more strongly with perceived threat to professional identity than with skill gaps. Employees who feel confident in their ability to learn the tool still resist it when they believe it diminishes their role, removes their judgment from decisions, or signals that leadership views their current contribution as replaceable.
Training addresses the wrong layer. The real barriers sit beneath the skills surface, in the psychology of how people relate to their work, their competence, and their place in the organisation.
The Three Drives That Determine Whether AI Sticks
Decades of organisational psychology research, from self-determination theory to the work of researchers like Amy Edmondson on psychological safety, point to three fundamental drives that shape how people respond to change in the workplace. Each one is directly relevant to AI adoption.
Identity
Professional identity is built over years. An experienced analyst who has spent a decade developing expertise in financial modelling does not simply welcome a tool that can produce comparable outputs in seconds. Intellectually, they may understand that the tool makes them more productive. Emotionally, they experience it as a devaluation of the skill that defines who they are at work.
This is not irrational. Identity threat is one of the most powerful psychological responses humans experience, and it operates largely outside conscious awareness. The employee may not articulate it as "this tool threatens my identity." They are more likely to express it as scepticism about the tool's accuracy, complaints about workflow disruption, or quiet disengagement from the adoption programme.
Leaders who dismiss these responses as resistance to change are missing the signal. The employee is not resisting the technology. They are protecting something that matters to them deeply.
Autonomy
The second drive is autonomy, the sense of control over how work gets done. AI tools, particularly those that automate decision-making or generate recommendations, can feel like a reduction in professional agency. The employee moves from being the person who decides to being the person who reviews what the machine decided.
For many knowledge workers, autonomy is a core part of what makes their job satisfying. Research published in the Journal of Applied Psychology consistently shows that perceived autonomy is among the strongest predictors of job satisfaction and intrinsic motivation. When AI adoption reduces that perception, even if the tool is genuinely helpful, the psychological cost can outweigh the productivity benefit.
The employees who resist most strongly are often the ones who were previously the most autonomous: senior individual contributors, subject matter experts, team leads who had developed their own methods. They are the people with the most to lose.
Trust
The third drive is trust, specifically the trust between employees and leadership about why AI is being introduced and what it means for their future.
When leadership frames AI as a tool for "efficiency" and "productivity," employees hear a subtext, whether intended or not. Efficiency gains in headcount-heavy functions have historically led to workforce reductions. Employees are not naive about this. They read the industry press. They know which companies have announced layoffs alongside AI investments.
Deloitte's 2025 Global Human Capital Trends report found that trust between employees and employers has declined in many industries over the past three years, precisely the period when AI adoption has accelerated most rapidly. The correlation is not coincidental. Organisations that deploy AI without transparently addressing what it means for roles, careers, and job security are spending trust they may not have in sufficient supply.
Why More Training Does Not Solve a Psychology Problem
Training is the default organisational response to adoption failures because it is visible, measurable, and comfortable. Leadership can point to it as evidence that they invested in preparing the workforce.
But training operates on the assumption that resistance comes from a lack of capability. If the real barriers are identity threat, autonomy loss, and trust erosion, then additional training sessions do nothing to address them. In some cases, they make things worse, because they signal to employees that leadership does not understand what is actually going on.
The research on technology adoption in organisations supports this. The original Technology Acceptance Model, developed by Fred Davis in the late 1980s and refined extensively since, identifies perceived usefulness and perceived ease of use as drivers of adoption. But later extensions of the model, including Venkatesh's Unified Theory of Acceptance and Use of Technology, incorporate social influence and facilitating conditions, recognising that adoption is a social and emotional process, not just a cognitive one.
Addressing the psychology of AI adoption requires a different set of interventions than training alone can provide.
What Actually Works
Organisations that successfully navigate AI adoption resistance tend to do three things differently.
First, they address identity proactively. That means framing AI explicitly as a tool that elevates the employee's role rather than diminishes it. Not in a slide deck, but in the actual design of how the tool is integrated into workflows. If the tool takes over the routine parts of the job, the organisation needs to simultaneously expand the employee's scope into higher-value work. The framing has to be backed by structural change, or it rings hollow.
Second, they protect autonomy. That means giving employees meaningful choice in how they use AI tools, rather than mandating adoption and measuring compliance. The most successful rollouts we have seen are the ones where employees are invited to experiment, given agency over how they integrate the technology into their personal workflow, and asked for feedback that visibly influences how the programme evolves.
Third, they build trust through transparency. That means having honest conversations about what AI means for the organisation's workforce. Not reassuring platitudes about "augmentation not replacement," but specific, credible commitments about how roles will evolve, how upskilling will be supported, and what the organisation will and will not do with the productivity gains AI produces.
The Culture Layer Comes First
Technology is in place in most organisations. The limiting factor is not the AI. It is the culture the AI is being deployed into.
If your organisation is experiencing resistance that more training has not solved, the answer is probably not in the technology or the curriculum. It is in the psychology.
We embed behavioural change methodology into every AI programme we run, addressing fear, resistance, and identity threat directly. If that is a conversation your leadership team needs to have, we are ready for it.
Talk to our team at kydongrp.com/contact
Sources: Gartner. "How to Overcome Digital Workplace Resistance, 2025." https://www.gartner.com/en/information-technology/topics/digital-workplace Deloitte. "2025 Global Human Capital Trends." https://www2.deloitte.com/us/en/insights/focus/human-capital-trends.html Venkatesh, V. et al. "User Acceptance of Information Technology: Toward a Unified View." MIS Quarterly, 2003. Edmondson, Amy. "The Fearless Organization." Wiley, 2018.