Employers are deploying AI. Workers are figuring it out alone. And everyone is calling it a strategy.
This notion has been sitting with me since we completed research into how workers are really experiencing the AI transition. Not the version organizations are communicating, but the lived reality on the ground. What we found wasn’t a story of cautious resistance or generational friction that tends to dominate the public conversation. It was something more quietly alarming …
A workforce that has been handed powerful tools, pointed in a general direction, and largely left to work it out on their own. It’s also why we built Cornerstone Workforce AI™ to provide a direct answer to what we kept seeing in the data. Organizations have a genuine intent to support their people as AI evolves how they work, yet there’s a notable structural gap between that focus and what workers experience.
Nearly half of the employees using AI tools at work (46%) say they have never received formal training for it. That alone should give any leadership team pause. But the more revealing story is in what workers did next. Without support, they didn’t stop. They adapted.
Nearly half taught themselves through trial and error. More than a third deliberately limit how much they use the AI tools at their disposal, rationing their exposure to manage the risk of getting it wrong. And one in six simply pretend to use technology they’ve never been equipped to use, in performance environments that apparently require the appearance of adoption.
From my standpoint, faking AI use isn’t apathy or resistance. It’s the end state of a worker who lacks formal guidance, has no definition of what good looks like, no psychological safety net to say they’re struggling, and an organization that has created the experience of AI use without creating the conditions for it.
The strategy that isn’t
Here’s what makes this harder to dismiss: Most organizations believe they’re handling this well. Our research revealed that 65% of employees believe their company has an AI upskilling strategy. In fact, 75% say leadership has clearly identified the skills the business needs. It seems like good progress, but is it?
When you push past the surface, only 33% say that strategy has actually been translated into training. Only 36% say it’s well communicated. Conversely, when you combine the 27% working in organizations with no formal strategy at all with the 29% whose strategy exists but hasn’t been well communicated, suddenly we’re looking at a majority of the AI-using workforce (56%) with no effective path to upskilling.
No clarity. No program. No map to success.
The gap between strategic objectives and organizational delivery is a structure and operating model failure, and the difference between an organization that has determined AI upskilling matters and one that has actually done the work of defining what that means, role by role, function by function, in terms that workers can act upon.
What workers already know
Here is the finding I keep coming back to, because I think it changes the conversation.
When we asked workers which AI skills will matter most to their careers in the long run, they didn’t gravitate first toward prompt writing or technical AI knowledge. They ranked critical thinking and judgment at the top, alongside other very human skills like creative thinking, problem-solving, resilience and adaptability. Technical skills were acknowledged but were clearly understood as surface capabilities rather than foundations.
These aren’t aspirational answers, but conclusions of workers who are using AI daily, navigating it mostly on their own, and drawing direct experiential conclusions about what actually makes it useful. What they’ve concluded is the capability that matters most is the judgment to evaluate what the tool produces, knowing when the output is good, when it should be questioned, and what to do when it falls short.
Workers have, in other words, worked out what’s missing (for now) through a messy, unstructured and unvalidated process. But as the pace of change continues, business models, technology and skills continue to evolve, will this be enough in 3 months or 6 months?
The actual work
Forget AI adoption as a technology problem. Focus on the organizational clarity challenge.
The organizations that navigate this well are the ones that have role-level clarity before program design, leadership behavior that matches communication, and genuine investment in the judgment and critical thinking workers have identified as the foundation of everything else.
This is precisely what Cornerstone Workforce AI™ is built to deliver. It gives organizations a clear, continuously updated picture of what their people can actually do. That’s not based on job titles or self-reported skills, but on real signals from how people work and learn, mapping deficiencies at the individual, team and organizational level. That means leaders can stop guessing at the gaps and make decisions based on what’s really happening. And employees? They get to stop operating on assumption because their expectations are specific and actionable.
We also connect that picture to what the business needs — and what the market is demanding — so the right development reaches the right people at the right time. When a people leader or business owner stands up and can say, “Here’s how we’re building AI capability across our organization,” the message lands differently when it’s grounded in workforce intelligence. Not aspiration or a deck, but actual data about what the workforce knows, where it’s falling short, and what’s being done about it.
That’s the difference between calling something a strategy and having one.
For organizations willing to build around that insight, the path forward means giving individual effort the structure, direction and conditions that let it compound into something shared.
