Most conversations about AI adoption begin with the wrong measurement. Instead, of counting licenses purchased, trainings completed, and dashboard activity, there’s a more effective way to know if your AI strategy is working. Unfortunately, the numbers coming out of 2026 suggest that proof of transformation is thinner than it looks.
A May 2026 survey of U.S. office workers found that 88% now use AI in their jobs. That is close to universal. Yet Gallup’s research shows only about one in ten employees strongly agree that AI has actually changed how work gets done inside their organization. Using AI at work is more and more popular, but the transformation everyone was promised has largely stayed on the slide decks.
There is a well-studied reason a gap like this forms, and it has very little to do with the technology.
Years ago, organizational psychologist Dr. Tasha Eurich set out to understand self-awareness. Her research found that roughly 95% of people believe they are self-aware, while only 10 to 15% actually are. The people living in that gap are not unintelligent. They simply can’t see themselves how others see them. Eurich describes self-awareness as two separate skills. Internal awareness is how clearly you see yourself. The external kind is about how accurately you understand the way others experience you. Most people have an idea about themselves, but the second one, they have no clue, and leaders are not exempt.
I want you to hold that finding up against an AI rollout, because the same blind spot is hiding inside it.
Internally, leadership sees all the evidence of progress. They see the tools they invested in, training sessions in progress, and the dashboard stays busy. Externally, the people doing the work with AI may be having a very different experience. While one person is checking every output twice before trusting it, another colleague nearby has quietly slipped back to the old way and only opens the tool when a manager walks past. Someone else on the same team is genuinely flying with it. All of that can be true at once, and the dashboard will still read as usage. The activity is real. What that activity actually means is where leaders can get misled.
That interpretation gap carries a cost. Last year, a METR randomized controlled trial gave experienced developers AI tools to use on real coding work. They expected to move 25% faster. In reality they came in around 19% slower, and they truly believed the tools had sped them up. Capable people, using the technology, confidently wrong about their own experience of it. That is Eurich’s research showing up inside a codebase.
So the real question is not whether your organization has adopted AI. Nearly everyone has. The question is whether you can see clearly how your people are actually working with it. And this is where even a genuinely self-aware leader can still come up short, because there is a third kind of awareness that internal and external reflection will never surface on their own.
In my world, it’s called conative awareness.
Conation is the part of the mind that governs how a person naturally takes action. It is distinct from the cognitive domain, which is about what a person knows, and from the affective or feeling domain, which is about what a person feels. Conation is the instinctive way each of us is driven to get things done, and it is remarkably stable over a lifetime. Give two people the same tool, the same training, and the same deadline, and they will still move through the work along different instincts. One begins by gathering the facts and verifying before committing. Another begins by testing something and adjusting on the fly. Both are effective, and each is simply wired to solve problems a particular way.
When a rollout quietly assumes everyone will adopt AI through the same instinct, it manufactures friction for every person whose wiring does not match the assumption. That friction shows up as a performance tax. It rarely shows up on a report, yet your people pay it in the small hesitations, the redundant double-checks, and the workarounds nobody mentions in a status meeting. Left unnamed, it looks like resistance or a training gap, when the real issue is design.
This is the missing link. Linking Dr. Eurich’s research connecting leaders to internal and external awareness as the foundation of sound judgment. AI adoption asks for one more layer on top of it. Conative intelligence is the ability to see how your people are naturally built to take action, and then to design the work around that reality rather than against it. When all three forms of awareness line up, AI stops being a tool you deployed and starts becoming an advantage you have actually earned.
So if your dashboard says the strategy is working, it is worth asking what that dashboard can and cannot see. A dashboard is very good at counting usage. It has no way of telling you whether your people are thriving inside the change or quietly absorbing the tax. For that answer, you have to understand how they are wired to act.
That understanding is available, it is measurable, and it reshapes how every rollout creates real transformation.
Alicia Couri is Founder and CEO of Audacious Concepts Inc.