There’s a version of AI adoption that looks like progress from the outside. The company has announced an AI initiative. Someone has been assigned to lead it. A tool has been selected and provisioned. And yet, six months in, usage is inconsistent, the team isn’t sure what they’re supposed to be doing with it, and the initiative has lost whatever momentum it had at the start.

In most of those cases, the problem isn’t the tool. It’s that leadership never actually got in the game.

The Do-As-I-Say Problem

Henry Hays works with business owners and executives across industries on AI adoption, and one of the first questions he asks when he sits down with a leadership team is simple: do you use it?

The number of senior leaders who say no is striking. And it creates a problem that no amount of internal communication or training can fully solve. When leadership isn’t using AI themselves, the message the organization receives, regardless of what gets said in all-hands meetings or strategy documents, is that this isn’t serious enough to change how the people at the top work.

Teams are perceptive. They watch what their leaders actually do, not what they say. If the CEO is telling the organization to embrace AI but isn’t visibly using it in their own work, the implicit signal is that AI is something for other people to figure out. And that signal travels fast.

The leaders who drive real adoption are the ones who can sit down with a team member who is stuck on an AI tool and actually help them. Not because they read an article about it, but because they’ve been using the tool themselves and they know what it can do. That kind of leadership credibility can’t be faked, and it can’t be delegated.

What Leadership Modeling Actually Looks Like

Becoming a practitioner doesn’t mean becoming an AI expert. It means using AI consistently enough in your own daily work that you develop a genuine feel for where it helps and where it doesn’t.

For most business owners, this starts with the parts of their workweek that are repetitive or time-consuming: preparing for meetings, drafting communications, summarizing information, thinking through decisions. These are the areas where AI tends to deliver quickly and clearly, and where a leader can build enough personal experience to speak about it authentically.

The key shift Hays describes is moving from treating AI like a vending machine to treating it like a real assistant. Pushing a button and expecting a packaged output is how most people start. The leaders who get real value out of it are the ones who learn to have a conversation with it, push back on what it produces, redirect it when it misses, and iterate until they get something useful. That’s a skill that develops through use, not through reading about it.

Once a leader has that experience, their relationship to the AI initiative in their organization changes. They’re not sponsoring something they don’t understand. They’re leading something they’ve actually done, and that difference shows up in every conversation they have about it.

The Ownership Problem on Lean Teams

One of the most common questions in companies without a dedicated technical team is some version of: who’s supposed to be running this?

The instinct is to point to IT. But IT ownership of an AI initiative, particularly in a growing business where the IT function is already stretched, often means the initiative doesn’t move. The people with the deepest technical knowledge aren’t always the people with the most to gain from AI, or the ones best positioned to drive adoption across the rest of the organization.

What Hays consistently finds is that the right person to own AI inside a company is whoever is already curious about it on their own time. The person who has been experimenting with tools on weekends. The person who brings up AI in team meetings because they’ve been thinking about it. The person who is excited enough about it that they’d work on it even without being asked.

That person might be in marketing. They might be in sales. They might be in operations. They almost certainly aren’t the person you’d pick if you were going by job title alone. And when you find them and give them organizational support and resources to go deeper, you end up with someone who can actually move the initiative forward because they’re intrinsically motivated to do so.

The jobs that will exist inside organizations to manage and govern AI don’t all have names yet. Some of the behaviors people are already doing under their current titles are the early version of roles that will be formalized over the next several years. Finding the people who are already doing those behaviors, and investing in them now, is one of the most practical things a business owner can do.

Giving the Team What It Needs to Move

Once leadership is in the game and ownership has been established, the next thing that has to be in place before adoption can scale is a clear set of guidelines for how the organization is going to use AI.

This doesn’t need to be a lengthy document. It needs to be a clear, practical answer to the questions every employee is going to have: what tools are we using, what can I do with them, what should I not put into them, and who do I go to if I have a question?

The reason this matters is that employees aren’t waiting for the company to figure out its AI strategy before they start using AI. They’re already using it. They’re bringing personal accounts to work, running work content through tools the company hasn’t reviewed, and making their own decisions about what’s appropriate because no one has told them otherwise.

A usage policy doesn’t have to be restrictive. The best ones send a clear message that the company is taking AI seriously, that it’s safe and encouraged to use it, and that here are the boundaries that protect both the employee and the business. That kind of clarity creates the conditions for real adoption, because people are much more likely to use something confidently when they understand the rules.

The Compounding Effect of Getting This Right

Leadership adoption and organizational clarity compound over time in a way that’s hard to see at the beginning but becomes obvious later.

When leaders use AI and talk about it openly, it normalizes the behavior for the rest of the organization. When there are clear guidelines, people experiment more freely because they’re not worried about doing something wrong. When the right internal owner is resourced and supported, the initiative has someone who is accountable for making it work and motivated to do so.

The businesses that are generating real returns from AI right now didn’t get there by buying the right tools. They got there because someone in leadership decided to take this seriously enough to model it, and that decision created the conditions for everything else to follow.

Listen to the full podcast episode here.

Related Resources

• The AI Execution Problem: Why Most Growing Businesses Can’t Get Past Experimenting: Leadership modeling is one piece of a larger pattern. This post covers the full picture of why AI execution breaks down in growing businesses and what it actually takes to get through it.
• Your Employees Are Already Using AI at Work — Here’s What That Means for Your Business: Employees aren’t waiting for the company to have an AI strategy before they start using AI. Here’s what unmanaged adoption looks like and why a usage policy matters more than most leaders realize.
• Who Should Own AI in Your Company When You Don’t Have an IT Department: Finding the right internal owner for AI isn’t about job titles. Here’s how to identify the person who will actually move the initiative forward and what to do once you’ve found them.
• The Boring, Mundane, and Tedious: Where AI Actually Delivers for Growing Businesses: Before leaders can model AI use, they need a place to start. This post identifies the workflows where AI delivers the fastest, clearest returns for business owners and their teams.