Whether Your Team Adopts AI Comes Down to Trust
Key Takeaways
Whether a team adopts a new tool has less to do with the rollout plan than with how safe people feel to learn it in real time in front of their leader.
AI is not a typical change. It touches competence, raises feelings of threat from technological advancement, and what people believe their contribution is worth. This gets to the core of our identity as humans, which raises the stakes of that safety considerably.
Safety isn’t created when a new tool shows up. It comes from how a leader has reacted, over months, every time someone got something wrong, raised a dissenting opinion, or experimented and failed.
This has happened before. The word “automation” was coined at Ford in 1947, and how leaders handled the human side of that transition determined how well they navigated it.
You can’t retrofit trust in a launch meeting. But you can start building it now before the next big change.
Every organization I talk to is asking some version of the same question: how do we get our people to actually use AI? Most are treating it as a training-and-rollout problem. Buy the licenses, book the workshops, send the launch email.
In my experience, it is mostly a trust issue, and whether or not you have the requisite trust is something that has already been decided long before AI came on the scene.
Why adoption suffers despite a solid rollout plan
Here is a pattern playing out in organizations everywhere right now. The rollout plan is airtight: training booked, licenses provisioned, a thoughtful change email sent. Six months later, half the team has reverted to the old way of working.
The explanation is rarely the tool. Adopting anything new means being visibly bad at it for a while, in front of people who evaluate you. People don’t weigh the tool’s merits in the abstract. They weigh what happened the last time they fumbled something in front of this leader.
The research supports this. Jamil Zaki, a Stanford psychologist who studies human connection, makes the case in Harvard Business Review that empathetic leadership can make or break AI adoption: employees with caring managers are far more likely to experiment with new tools and take the risks that adoption requires.
AI is not a typical change
It is worth being honest about why this particular change feels different. It’s because it is different.
Most changes alter a process. And while AI does change our processes, it does something more. It changes our relationship to the work itself and to each other. It touches identity, competence, and what people believe their contribution is worth. People hear daily that AI may take their jobs, and the noise keeps raising the temperature. A person who suspects a tool threatens their role is not going to champion its adoption, no matter how good the training is.
People still move along the familiar change curve: the stages of denial, resistance, exploration, commitment. But this curve is accelerated and more chaotic, with more unknowns and less precedent to keep a steady course.
Which makes this a change-management challenge before it is a technology challenge, and change management has a humbling track record. McKinsey’s transformation practice puts the failure rate of large-scale change efforts around 70 percent. Organizations will adapt to AI with widely varying success, and the difference will not be the quality of their tools.
When change stalls, leaders tend to reach for “optimization”. They tinker with tuning the rollout, where to add value, offer more training, and provide more support. The more durable answer sits underneath: the core principles of trust, carried through everything from how the decision is made to how it is communicated and followed through.
Your team already decided how safe it is to make mistakes, fail, or learn
The safety calculation your people are making right now is this: Is it okay to be clumsy with this in front of my leader? They have already decided the answer to this question based on their previous experience in a similar situation.
I have written before that trust is built in small moments, not big announcements. AI adoption is that principle meeting its newest test. So “are we AI-ready?” is really a question about the past. What have people learned, over time, about what happens when they try something and it doesn’t work?
This is also why the workshop approach falls short here, just as it does with psychological safety more broadly. Safety is not a concept people need explained. It is an experience they need to have had, repeatedly, with their actual leader.
Why this gets more true as the technology advances
The more advanced the tools become, the more the deciding factor becomes about our human qualities.
As AI absorbs more analytical work, the scarce leadership capability is how you hold uncertainty, how you respond when someone is wrong, how you make it possible for people to say “I don’t know what I’m doing yet” or even “I don’t know if I trust this new technology.” The instinct in most organizations is to treat AI as a skills problem for the team. It is at least as much a behavior question for the leader.
I explored this territory in Leading From the Inside Out. The human capabilities that matter most are the ones that shape how others experience us, and those capabilities scale in importance as everything else gets automated.
This has happened before
We’ve seen this movie before, as it were. And the plot is instructive for our current purposes. Let’s take a look.
The word “automation” was actually coined inside a company: Ford. In April 1947, manufacturing vice president Delmar Harder created Ford’s Automation Department and the word spread from there. In fact, Harder came to be known as “the father of automation”. His department started with five people. Within a year it employed fifty, and it was automating the manufacture of pistons, wheels, and axles.
What followed reshaped manufacturing: the work, the roles, the fears about jobs, and the relationship between people and the production line. Whole categories of work changed, and whole categories emerged. The anxieties in the headlines then would read as familiar today.
The AI moment has a similar shape with different details. It’s a general-purpose capability arriving faster than the norms for using it, met with a mix of possibility and fear and the polarizing public discourse that comes along with that. And the lesson from that history is not that everything works out. It is how organizations and their leaders handle the human side, not the machinery, which ultimately determines who comes through it well.
What to do before the next change comes on the horizon
When you are already on the precipice of change, it’s too late to manufacture trust. But there is hope for the future. What you can do is start building it now, in the ordinary moments that precede the next big change.
Audit the reputation you are accumulating. When someone fumbles in front of you, what do they learn about whether to try again? When someone says “I don’t know,” what happens next? Those responses are setting your team’s willingness to take risks, including the risk of adopting the tools you need them to adopt.
This is where coaching does some of its most practical work: making those micro-reactions visible to the leader producing them, before the next change arrives to test them.
FAQs
Isn’t AI adoption mostly about training and the right tools?
Training and tools are necessary, but they explain surprisingly little of the variation in whether teams actually adopt. Two teams can receive the same rollout and land in very different places — and the difference usually traces to whether people feel safe being visibly inexperienced in front of their leader. Training addresses capability. It doesn’t address whether anyone is willing to look like a beginner.
What does trust have to do with whether people use software?
Using a new tool means being bad at something for a while, in view of people who evaluate you. That is a risk calculation before it is a technical one. If past experience says mistakes on this team are met with curiosity, people experiment. If it says mistakes are remembered, they keep using what they already know, and it’s reflected in your adoption metrics.
Can I build this kind of trust quickly if a big rollout is coming?
Not on the rollout’s timeline, honestly. The safety people feel is the product of months of accumulated interactions, and a launch meeting can’t replace that history. What you can do immediately is change how you respond from today forward. Treat early fumbles with the new tools as expected rather than concerning, and be visibly willing to struggle with them yourself. That won’t rebuild years overnight, but teams notice a genuine shift faster than you’d expect.
How do I know if my team feels safe to experiment?
Watch what happens around mistakes and questions. Do people surface problems early, or do issues arrive fully formed and late? Do they ask questions in the meeting, or afterward, privately, one-on-one? Do they show you rough drafts, or only finished work? A team that only brings you polished output is telling you something about what it costs to be unfinished in front of you.
Closing Reflection
Think about the last time someone on your team got something wrong in front of you. What did they learn about whether it’s safe to try the next new thing?
Explore More
→ Trust Is Built in Small Moments, Not Big Announcements
→ Psychological Safety Isn’t Built in a Workshop
→ What Leaders Communicate Without Saying a Word
→ Leading From the Inside Out: Where Emotional and Social Intelligence Meet
Download
The relational gap, or how others experience you, is one of the three gaps in my Leadership Patterns: A Reflection Workbook, and it maps directly onto whether people will take risks in front of you. It’s free to download.