gshc2020.com

A leadership team I worked with recently did everything right on paper: licenses bought, kickoff held, people walking out excited. Months later, most had drifted back, and not one blamed the software. What I heard was "I didn't trust it enough to put my name on what it gave me."

I have spent more than two decades at Future Point of View helping organizations through big technology shifts. AI is different: It does not just ask people to learn a tool; it asks them to rethink how they work, and most companies still treat it like a software rollout.

The Mistake Most Companies Make

Almost everyone starts with training. Here is the platform, here are some prompts. What they skip is what makes it stick: helping people grasp what the technology does well, where it falls down, and where their own judgment still carries the day. Skip that and people either trust the tool too much or avoid it altogether.

There is a version of this that starts in the corner office. Leaders hand AI off to IT as a technical problem when it is really a workforce-design decision. Move faster than people can absorb, and the gap fills with anxiety, foot-dragging and the nobody-is-watching experimentation that should worry you most.

A Framework For Building That Understanding

The employees who adapt best are rarely the most technically proficient; they are the ones who have honestly worked out what AI can and cannot do in their own jobs. We start every engagement with three questions everyone should be able to answer about their own role.

First, which parts of my work are really just pattern recognition? AI is good at volume and repetition: first drafts, summaries, sorting feedback. Naming those tasks drains the fear: People stop seeing a replacement and start seeing room to breathe.

Second, where does my work depend on things AI cannot know? The history of a relationship, the politics in the room, the thing a client said that was never in an email. The people who use these tools well carry that context in rather than taking the output at face value.

Third, what does good actually look like in my world? A seasoned person can glance at an AI draft and feel what is thin or overstated, what would collapse the moment a client pushed on it. That instinct is the muscle organizations forget to train.

I watched this land on a client-services manager who came in with her arms crossed. By the first question, she admitted that the weekly summaries she dreaded were pure pattern; by the third, she was naming the judgment that made the job hers rather than asking whether the tool could replace it. That is what I mean by the new Humalogical Balance: human capability, tools and automation lined up on purpose so judgment leads and technology serves.

What This Looked Like In Practice

One engagement stays with me. A regional organization of a few hundred people had hit exactly that wall. Rather than repeat the training, we got honest about where each team stood, built capability at a manageable pace, reworked broken workflows and let a few roles change openly. The hardest resistance came from a respected veteran who took it as an insult to his experience. The turn came when we asked him not to adopt the tool but to tear it apart, holding his judgment against its output. He went from the biggest obstacle to the loudest advocate, and flat adoption began to move.

The Comfort Question Is Actually A Trust Question

Leaders love to file resistance under comfort, assuming time will settle the nerves. In my experience, that is half of it. The rest is trust, running both ways: People need to believe you are not quietly using AI to thin the payroll, and to feel safe challenging a tool that can be confidently, articulately wrong.

I think of a junior analyst who caught a polished AI summary that had flipped a key number backward. In many companies, she would have stayed quiet. She spoke up, and her manager thanked her in front of everyone and made clear that catching it was the job now. You could feel the room relax. People take their cue from what happens to the first person brave enough to speak up.

None of this is soft. When flagging an error feels like admitting you are behind, people nod along with bad output or use the tool where no one can see. The organizations handling this well give explicit permission: to experiment, to say when the output misses, to build habits that fit the real work rather than a tidy version on a slide.

What Leaders Need To Do Differently

The executives moving the needle use the tools themselves, visibly, which says more than any memo. They make space for teams to talk openly about what AI gets right and wrong, and treat it as ongoing rather than a box checked at a Tuesday training. What lasts is the thinking underneath: assessing, adapting and applying judgment to whatever comes next. That is how you move people from fear-based friction to real momentum, long after the launch buzz fades.

Getting your workforce comfortable with AI is not a technology problem, or even a training problem. It is change management, and it requires the same honesty and patience that every real transformation demands. The strongest AI-empowered workforces will be built by organizations that take the time to understand AI deeply instead of simply rushing to adopt it. Success depends on more than whether your people are using AI. It depends on whether they understand it well enough to use it effectively, and whether your organization has created an environment where that understanding can grow.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?