My husband has a saying: nobody likes change.
He isn’t entirely wrong. But I love change and find introducing it at work energizing. I’ve always been curious about why some people lean in while others dig in.
Building my systems-thinking “muscles” changed how I understand that gap. Processes, structures, and behaviours reflect underlying beliefs. When change stalls, the useful question is not only, “What’s wrong with the rollout?” but “What belief is this bumping up against?”
I didn’t have that frame fully formed when we introduced an enterprise project-management tool a few years ago. We needed visibility, consistency, and an end to the spreadsheets and status emails consuming everyone’s time.
The response surprised us.
Looking back, I think some project managers experienced the tool as criticism, staff saw a route to micromanagement, and leaders saw another system to learn. Nobody said that directly. We heard: the tool is too complicated. The timing is bad. We don’t have the processes in place. Can we delay the rollout?
They were right about the processes. We had skipped foundational work, creating rework and frustration. But that did not explain the whole response. Process was simply the safest concern to name.
It took me time to see what else may have been underneath. I think many organizations are now encountering the same pattern with AI.
Canada has a significant AI adoption gap. For every 10 small and medium-sized Canadian businesses, fewer than 1 has adopted AI. In France, it’s nearly 2 in 10. In Germany, roughly 1 in 4. Among Nordic leaders, it’s closer to 3 – 4 in 10. Put another way: a Nordic SME is roughly four to five times as likely to have adopted AI as a Canadian one. (Source: Government of Canada, 2026 National Artificial Intelligence Strategy.)
That gap is not simply about access to technology. Mercer’s 2025 Canadian employee research found that for every 2 employees who say new technology makes them more efficient, roughly 1 finds it frustrating or hard to use, and nearly half are worried it will cost them their job.
What AI Reveals
The pattern is older than AI. Enterprise software, ERP systems, and performance dashboards all changed how work was done and made individual capability more visible. Each wave produced resistance that was not only about the tool.
Most professionals have moments when their confidence exceeds their certainty. Familiar work allows that gap to remain private. A new tool can expose it. Suddenly, the distance between what you know and what you are expected to know is visible to others. The fear is not only being replaced, it is being revealed.
Brené Brown’s research helps explain why this experience can feel so personal. In Dare to Lead, she distinguishes guilt — “I did something bad” — from shame: “I am bad.” Shame is not simply discomfort about making a mistake. It is the fear that the mistake reveals something fundamentally inadequate about us.
Brown also identifies self-worth tied to productivity as one way shame can show up at work. That connection matters when experienced professionals are suddenly asked to become beginners. If competence has become part of how someone understands their value, struggling with a new tool may feel like more than a learning curve. It may feel like exposure.
Brown argues that honest conversations about shame require the right conditions: people need to feel safe enough to name what is happening. Without that safety, concerns about competence may surface indirectly — as objections to the process, the timing, the training, or the tool itself.
The implication for AI adoption is straightforward: people will not admit what they do not know unless it feels safe to do so. AI adoption asks experienced professionals to become beginners in public—to ask basic questions, work more slowly, and produce imperfect results while they learn. If doing that carries a professional penalty, people will protect themselves. If people cannot learn visibly without risking how their competence is judged, AI adoption will carry more risk than leaders may realize.
Three Places to Look Before You Push Harder
None of this can repair a performance system or culture that has been moving in the wrong direction for years. But when AI adoption stalls, the answer is not always more training, better communication, or stronger expectations.
Before pushing harder, look at what people may believe they are losing, what it costs them to learn, and what your systems are telling them to do.
1. What People Believe They May Lose
Resistance is often described as fear of change. That is rarely specific enough to be useful.
People may worry about losing expertise they have spent years building. They may see their influence shrinking as knowledge becomes easier to access, fear being exposed as less capable than others assumed, or wonder whether automation will make parts of their contribution less valuable.
Those concerns may appear as criticism of the tool, repeated requests for more information, insistence that the work is too unique, or reasons why now is not the right time.
Ask:
What does this change threaten to take away—and have we acknowledged that loss before labelling the response as resistance?
2. What It Costs to Be a Beginner
Leaders often encourage experimentation while continuing to reward speed, certainty, and polished performance.
That creates an impossible expectation: learn something genuinely new, but do not slow down, make visible mistakes, ask basic questions, or produce anything below your usual standard while doing it.
Watch what happens when someone gets an imperfect result from AI. Do they receive help improving it, or does the mistake become evidence that they cannot be trusted with the tool? Can employees admit that they do not understand something, or are they expected to catch up quietly on their own time?
Ask:
What does it cost someone here to be visibly new at something—and what time, coaching, or protection would make learning possible?
3. What Your Systems Are Signalling
An organization can say that experimentation matters while its systems communicate something different.
Performance measures may reward individual expertise over shared learning.
Workloads may leave no room to practise.
Approval processes may punish initiative.
Leaders may encourage people to explore AI while also warning them not to make mistakes, create risk, or reduce short-term productivity.
When those signals conflict, employees will usually trust the system more than the message.
Ask:
What are our workloads, performance expectations, approval processes, and leadership responses teaching people to do?
Then align those systems with the experimentation and learning you say you want.
These are not abstract culture questions. They are operating conditions leaders can examine and change. Before asking people to adopt AI faster, leaders need to understand what the change threatens, lower the cost of learning, and ensure their systems support the behaviour they are requesting.
What Leaders Should Do Next
The project-management tool eventually became part of how we worked. Two years later, it was no longer controversial. The processes had improved, people had found their footing, and the visibility that once felt threatening had become useful.
However, we made the transition harder than it needed to be because we led with the tool. We assumed that a clear business case would resolve the resistance. What shifted was not another explanation of the technology. It was our willingness to acknowledge what we had underestimated and make the discomfort of not knowing safer to say out loud.
That experience changed how I think about adoption.
When people comply on the surface while withdrawing underneath, leaders should not assume the problem is simply resistance to technology. The change may be exposing concerns that were already present: what people believe they may lose, how safe it feels to be inexperienced, and what the organization’s systems reward.
Before pushing adoption harder, leaders need to examine four things:
- What are people being asked to risk?
- What does it cost them to learn in public?
- What behaviours do workloads, performance measures, and leadership responses reward or punish?
- What would make experimentation both safe and accountable?
AI does not only change how work gets done. It reveals the conditions surrounding the work — and successful adoption requires clearer expectations, stronger governance, and space for people to learn without hiding what they don’t yet know.
REFERENCES
Brené Brown, Dare to Lead: Brave Work. Tough Conversations. Whole Hearts. (2018), pp. 127–134.
Innovation, Science and Economic Development Canada. (2026, June 8). Canada’s national artificial intelligence strategy: AI for all. https://www.ised-isde.canada.ca/site/ised/en/canadas-national-artificial-intelligence-strategy-ai-all
Mercer, “Best Employers in Canada — Powered by Mercer,” citing Inside Employees’ Minds 2025 Canadian research. https://www.mercer.com/en-ca/insights/talent-and-transformation/best-employers-in-canada-2025-powered-by-mercer/
