Walk into almost any executive meeting about artificial intelligence and the conversation sounds familiar. Should we approve ChatGPT? Do we need an AI policy? Which enterprise platform should we buy? How do we govern employee use?
These are important questions. They are also increasingly the wrong place to begin.
I believe we’ve misunderstood where enterprise AI adoption is actually happening. The prevailing assumption is that organizations adopt technology from the top down. Leadership evaluates vendors, approves budgets, develops governance, trains employees, and then deployment begins. That is how enterprise software has traditionally entered organizations.
Consumer AI is different. I don’t think the first wave of enterprise AI adoption began in the boardroom. I think it began quietly at individual desks.
Picture the most junior employee in a company. They’re carrying a workload that keeps expanding while the hours in the day remain fixed. Their inbox never reaches zero. Deadlines keep moving closer. They want to leave the office in time to coach a Little League game, make dinner with their family, or simply get home before exhaustion takes over. They’re not looking for artificial intelligence because they’re fascinated by the technology. They’re looking for relief.
So they open a browser, ask their personal ChatGPT to summarize a document, draft an email, organize meeting notes, or help them untangle a spreadsheet. No procurement process. No steering committee. No executive approval. Just a practical solution to a practical problem.
Multiply that moment by millions of employees across thousands of organizations, and something remarkable begins to emerge.
I call it the AI Groundswell.
The AI Groundswell is the quiet, bottom-up adoption of consumer AI inside organizations before formal enterprise implementation. It isn’t driven by strategy. It isn’t driven by policy. It’s driven by human beings trying to reclaim an hour of their day. And it’s largely kept a secret.
Meanwhile, leadership is often having an entirely different conversation. Executives are evaluating enterprise platforms, debating governance frameworks, drafting acceptable use policies, and deciding whether AI should be introduced into the organization. Those conversations matter. Governance matters. Risk matters. But by the time many organizations begin discussing adoption, adoption has already happened.
The employees aren’t trying to break the rules.
They’re trying to solve a problem. That distinction matters.
Too often, organizations view unauthorized AI use as a compliance issue. I think it should first be viewed as organizational intelligence. If employees independently discover the same tools without being instructed to do so, they’re revealing where work has become unnecessarily difficult. Every hidden workflow points to friction. Every unofficial AI process points to unmet demand. Every personal AI account used to finish work more efficiently tells leaders something about how work is actually getting done.
After more than thirty years as an appellate lawyer, I’ve learned that behavior is often more revealing than policy. Courts don’t decide cases based solely on what people said they intended to do. They look carefully at what people actually did. Organizations should approach AI the same way.
Instead of asking only, “How do we stop unauthorized AI use?” leaders should also ask, “What problem are employees solving that our organization hasn’t solved for them?”
The answers to that question may prove far more valuable than the policy itself.
None of this diminishes the need for governance. Quite the opposite. As AI becomes woven into daily work, organizations will need thoughtful policies around confidentiality, security, intellectual property, and responsible use. Those issues are real, and they deserve serious attention.
But governance works best when it reflects reality. The reality is that consumer AI is moving at the speed of individual need, while enterprise governance moves at the speed of organizational process. Those two forces are on a collision course, and understanding that gap may be one of the defining leadership challenges of the next decade.
The organizations that succeed won’t simply be the ones with the best AI policies. They’ll be the ones that recognize the AI Groundswell for what it is: not an act of rebellion, but a signal. A signal that employees are searching for leverage, that work has fundamentally changed, and that the future of enterprise AI is being shaped from the bottom up, one prompt at a time.