Why is the C-suite excited about AI while the team is quietly terrified?

You walk out of the AI town hall feeling energised, and you can see in the room that almost nobody else does. You said all the right things. The slides were good. You meant every word about opportunity and augmentation and how this frees people for the work that matters. And yet the faces stayed flat, the questions were procedural, and the applause had the temperature of a fire drill. You cannot quite name what you felt in that room, only that your reading of it and the room's reading of it were two different events.

Here is the short answer to what you felt: you were standing inside the leadership-employee AI anxiety gap, and it is wider than almost any leader thinks. You are excited because you see the upside. They are quiet because they are doing private maths about their own role. The rollout is fine. The unaddressed fear is what stalls it. That gap is a coherence problem, by which I mean a mismatch between the state at the top of the house and the state on the floor, and until you name it, it silently throttles every adoption number you are trying to move.

Why does the room feel so much less excited than I do?

Because you are reading your own state and assuming it is theirs. The data on this is almost uncomfortable reading. Harvard Business Review surveyed 1,400 US employees and found that 76% of executives believed their people felt enthusiastic about adopting AI, while only 31% of individual contributors actually expressed that enthusiasm. Leaders were more than twice off the mark, reading the room in their own image. Derek Snyder of Google Workspace put the floor's posture plainly in a separate piece for CIO: employees "see the hype, but they're skeptical or dubious that they're going to be able to get real things done."

And the worry is not a vibe. Pew Research Centre, surveying 5,273 employed US adults, found 52% feel worried about future AI use in the workplace against just 36% who feel hopeful, with a third reporting they feel overwhelmed and about a third expecting fewer job opportunities for themselves over the long run. So the room you read as neutral was not neutral. It was holding something. The view from the bottom up is simply less sunny than the view from the stage, and your energy, sincere as it was, read to them as the energy of someone whose job is safe.

Is the fear actually about the technology, or about something else?

It is about meaning, and meaning lives below the tool. When you ask someone to start using AI for the core of their work, a quiet second message travels underneath the request: the organisation is preparing for a version of this role that needs fewer of you. SHRM (the Society for Human Resource Management, the largest professional body for HR in the US) names what follows precisely. It "shows up as reluctance to experiment, half-hearted adoption, or passive resistance." Nobody refuses. Everyone complies. And the energy leaks out of the whole thing.

You can see how private the fear has become in how people hide the tool itself. Slack's Workforce Index, surveying 17,372 desk workers across fifteen countries, found that 48%, nearly half, would feel uncomfortable telling their manager they had used AI for a common task. The reasons they gave name the dread directly: 47% felt that using AI counted as cheating, 46% feared looking less competent, and 46% feared looking lazy. People are not refusing the technology in the open. They are quietly worried that reaching for it marks them as replaceable, so they reach in secret or barely reach at all.

This is the part that surprises leaders most. The fear is not confined to the floor, and it is not evenly distributed by seniority in the way the org chart would predict. BCG, surveying 10,635 workers across eleven countries, found leaders and managers (43%) are actually more likely to worry about losing their job in the next ten years than frontline employees (36%), and that people inside organisations going through comprehensive AI-driven redesign worry more about job security (46%) than those at less-advanced companies (34%). The closer the change gets to real, the more the anxiety climbs. Same room, two emotional realities, and the gap between them is the thing throttling your adoption.

Look at the last line. EY surveyed 15,000 employees and 1,500 employers across twenty-nine countries and found that companies are forfeiting up to 40% of AI's potential productivity gains. The cause they identify is not the technology. It is what they call fragile talent foundations: weak culture, thin learning, rewards that pull the wrong way. In their data 88% of employees already use AI daily, yet only 5% use it in advanced ways, and just 12% report enough training. The capability is in the building already. What stalls is everything human around the tool. The bottleneck is no longer the technology.

What kind of intervention actually moves this?

Not another tools rollout. The intervention is coherence work, by which I mean deliberately closing the distance between what the top of the house feels and what the floor feels, so the two are aligned enough to move together. EY's number explains why this is the lever rather than a nicety: when the people foundations are fragile, up to two-fifths of the value simply evaporates, whatever the model can do. Most organisations have inverted the spend. They have poured budget into the software and left the human layer to look after itself, then wondered why the pilot went quiet.

There is a deeper reason the human layer governs the outcome, and it sits in the body. The neurovisceral integration research (work linking heart-rhythm patterns to the brain circuits behind self-control), indexed by heart rate variability, ties the integrity of the prefrontal circuits that govern self-regulation and clear executive judgement to cardiac vagal tone: higher vagal tone, better regulation and decision-making under load. A frightened nervous system experiments, learns and judges poorly. We see how fragile judgement is under depletion in the parole study of eight Israeli judges across more than 1,000 decisions, where favourable rulings ran near 65% early in a session and fell toward zero by its end before recovering after a break. (Later commentators have contested the size of that effect, and fairly.) The point holds at human scale: a workforce sitting in low-grade fear is a workforce making worse calls about the very technology you need them to adopt well.

So the work is to build coherence on purpose, in this order:

1. Name the split out loud. Say in the room what the data already shows: that excitement at the top and worry on the floor can both be true at once. Naming it is what lets people stop performing enthusiasm and start telling you the truth.
2. Answer the role question before the tool question. People decide whether to lean in based on what AI means for their future, so address that future first, concretely, before any feature.
3. Invest in the people foundations. Put real effort into workflows, governance and training, the layer where EY locates most of the lost value, rather than treating training as a single day that is meant to take and then does not.
4. Upgrade yourself first. Your own coherence, the alignment between what you feel and what you signal, is the field everyone else reads. The leaders who win the next decade are the ones who upgrade themselves first.

This is the quiet centre of the conscious approach to transformation, and it is where our work lives: technology's highest purpose is to serve human life, which means the human operating system, not the software, is the thing to tune. The town hall can stay as it is. The move is to get the energy at the top and the energy on the floor pointing the same way.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.