The Emotional Stack

I have started counting how long it takes, in a typical AI deployment conversation with leadership, before anyone mentions how their team is feeling about the change. The answer is almost never. The answer is almost always: zero people, zero minutes, never came up.

Then I spend half a day with the actual team, and I learn that the rollout is being privately resisted, undermined, or quietly ignored because the team is — collectively, often unconsciously — terrified. Of being replaced. Of being exposed. Of losing the skill that defines them. Of being asked to use a tool they don't understand and looking stupid.

None of this was on the deployment plan. It is the entire thing the deployment plan needed to be about.

It is helpful, I think, to be honest about what is actually being asked of a team when AI lands on their desk.

They are being asked to use a tool they did not select. To learn a new way of doing a job they have, in many cases, spent years mastering. To accept that a piece of their craft — the part they were proud of, the part that defined them — is now done partly or wholly by something else. To trust the output of a system whose mistakes they cannot fully predict. To do this quickly, on top of their existing workload, with no clear guarantee that the time investment will pay off for them personally.

And — crucially — they are being asked to do this in an environment where there is a background hum of jobs are being eliminated by AI. They cannot help but notice that the technology you are asking them to adopt is the same technology being credited, in the trade press, with making people redundant.

If you do not address any of this, you are running an emotional headwind so strong that no amount of training or tooling will overcome it. The technology layer is the tip of the iceberg. The emotional layer is the iceberg.

Every person on a team being asked to adopt AI is running these four questions in the back of their mind, whether or not they articulate them.

Am I safe? Is this rollout going to result in me losing my job? Is the speed of automation outrunning the speed at which I can stay relevant? If the answer is uncertain, the user will hedge — will use the tool to look compliant while keeping the old workflow alive in parallel. You will see adoption metrics that flatter you, while real adoption never happens.

Am I capable? Can I, personally, do this? Will I look foolish in front of my peers as I learn? Most people would rather not try than try and fail visibly. If the tool is hard, or the support is thin, or the room for error is small, users will avoid the tool to protect their professional dignity. They will not tell you this. They will say they're busy.

Do I have a future here? Even if I'm safe and capable, is the role I'm growing into a role I want? Is this organisation a place where someone with my new, AI-augmented capabilities is going to be rewarded? If the answer is no, your most capable people will adopt the AI rapidly and then leave with their new skills. You will mistake their adoption for success right up until they hand in notice.

Does my leadership see me? This is the deepest one and the one that's most often missed. Through this change, am I being treated as a person, or as a resistance-to-be-overcome? The answer to this question shapes every other answer. People who feel seen will work with you. People who feel managed-around will work against you, quietly and effectively.

The leaders I see successfully running AI transitions have a pattern, and once you see it, you cannot un-see it.

They name the discomfort first. They start every conversation about AI with an acknowledgement that this is hard, this is unsettling, and that the discomfort is reasonable. They do not start with the productivity gains. They start with I know this is a lot.

They commit publicly to the relationship. They tell the team, explicitly, what the deal is. "We are going to invest in you learning this. You are not going to be replaced for adopting this. Your job is going to change, and we are going to support you through that change." If they cannot make that commitment honestly, they don't make it — but then they don't pretend either, which is worse.

They make the learning visible and safe. They create explicit forums where it's okay to be confused, to ask basic questions, to admit you don't know what you're doing. They model this themselves. Senior leaders are first in the room asking the dumb question, because their visible vulnerability is what gives the rest of the team permission.

They track feelings as a metric, not just outputs. They check in regularly on how the team is feeling about the change. They take the feedback seriously. They adjust. They demonstrate that the feelings are taken as data, not as noise.

It can feel, when you look at the emotional work above, that this is a slow approach. That it would be faster to just push the tool out and demand adoption. That feelings are a luxury you don't have time for in a quarterly cycle.

I want to argue against that as forcefully as I can. The teams that handle the emotional stack with care are the fastest adopters I see. By a wide margin. The reason is that adoption is not a step you do after deployment. Adoption is something that either happens or doesn't, inside each person, when they decide for themselves whether to trust the tool, the rollout, and you.

If you have not earned that trust, no amount of technology investment will land. You will pay more, take longer, and end up with a worse outcome than the leader who spent the first month listening, naming, committing, and then moved at twice the speed thereafter because the team was actually with them.

Emotional intelligence is not the soft part of an AI transition. It is the foundation. Every part of the technical work rests on it, and the leaders who get this right are running circles around the leaders who think it's beneath them.