If the owner won't use AI, the team won't either

Plenty of owners want their team to use AI without using it themselves. They will buy the licences, send the encouraging message, even book the training, while quietly keeping their own work AI-free, on the grounds that they are too busy, or too senior, or that it is a thing for the younger staff. Then they are puzzled when adoption stalls. The puzzle solves itself the moment you accept how a small firm actually works. In a business of a dozen people, the founder sets the weather, and the team reads that weather far more accurately than any memo.

If the person at the top treats AI as something for other people, everyone else files it under optional. Not because they are told to, but because they watch what you do, not what you say. Greg Shove, who writes about AI and leadership, is direct about the obligation this creates: "we should be modeling this behavior for our teams, for our colleagues, for our friends and family." In a small firm the owner is the model whether they want to be or not. The only choice is what you are modelling.

The founder sets the weather

Culture in a small firm is not a poster on the wall. It is the accumulated example of what the people in charge actually do. Everyone takes their cues from the founder: what gets attention, what gets ignored, what is genuinely valued versus merely endorsed. This is the quiet superpower of a small team, and it cuts both ways.

Point it at AI and the logic is unavoidable. If you never mention using these tools, never reference them in your own work, never show a result you got from them, the team concludes that AI is talk, not practice. If you reference it naturally, the way you might reference any tool you rely on, it becomes part of how the firm works. You do not need a programme to set this weather. You need to be visibly in it.

Why delegated curiosity fails

The instinct to delegate AI to a keen junior and stay out of it yourself is understandable and a mistake. It fails for two reasons. First, it signals exactly the optionality you are trying to avoid: if it mattered, the thinking goes, the boss would be doing it. Second, and more practically, you cannot lead what you do not understand. Decisions about which tools to back, what to trust them with, where the risks sit, all require enough first-hand feel to judge well, and that feel only comes from use.

Jacob Morgan, who writes on leadership, makes a related point worth holding onto: most organisations are earlier in their real AI adoption than the noise suggests, and honest leadership matters more than performed confidence. You do not have to pretend to be an expert. You do have to be genuinely in it, learning alongside your team rather than directing from a distance you cannot see clearly from.

A 30-minute-a-week owner habit

Getting hands-on does not require a sabbatical. Half an hour a week, protected and regular, is enough to change what you model. The point is consistency, not volume.

Spend that half hour using AI on a real piece of your own work, not a toy exercise. Draft something with it. Think a decision through with it. Ask it to pull apart a problem you are stuck on. The work being real matters, because real work is where you learn what the tools do well and badly, and real work is what your team will see you doing. A founder fiddling with novelty prompts models novelty. A founder using AI on the actual business models the actual habit.

Your team will copy what you do for thirty minutes a week long before they act on what you say in a meeting.

Keep a light note of what you tried and what happened, partly to track your own learning and partly because it gives you something concrete to share. Over a few weeks, thirty honest minutes a week turns a sceptical or distant owner into one who can speak about AI from experience, which is the only kind of speaking the team actually believes.

Showing your working to the team

The habit only sets the weather if it is visible. So make it so, lightly and naturally. In the normal rhythm of the week, mention the thing AI helped with, and the thing it got wrong. Share the prompt that worked. Show the draft before and after. Say plainly when a tool saved you an hour and when it wasted twenty minutes.

The honesty is the important part. A founder who only ever reports AI triumphs models hype, which a smart team discounts. A founder who shares the real picture, useful here, useless there, confidently wrong on that, models judgement, which is exactly what you want the team to develop. Showing your working, failures included, gives people both permission to try and a realistic sense of what to expect.

The honest limits

Enthusiasm without guardrails is its own risk. An owner who dives in and starts feeding sensitive information into any tool that will take it, or trusting outputs without checking, is modelling the wrong habit just as surely as the absent owner is. The behaviour to model is thoughtful use: useful, but checked, and careful about what you share. Get visibly excited and visibly careful at the same time, because the team will copy both.

There is also a balance to strike on time. Thirty minutes a week is a habit; turning yourself into the firm's full-time AI tinkerer is a distraction from the job of running the business. The aim is to understand these tools well enough to lead their use, not to become the most prolific user in the building. Model the habit, then get back to leading.

A last point worth naming: you will be bad at this for a while, and that is fine to show. Founders often stay away from new tools precisely because they dislike being a beginner in front of their own team. But a leader visibly learning, fumbling a prompt, asking a junior how they got a better result, sets a far healthier weather than a leader who avoids the thing to protect an image of competence. Permission to be a learner is one of the most useful things you can model, because it is the permission everyone else is waiting for.

What to do this week

Block thirty minutes in the diary, recurring, and spend it using AI on a real piece of your own work. Keep a one-line note of what you tried and how it went. Then, in your next team conversation, mention it: what helped, what did not, what surprised you. Do that for a month before you worry about anyone else's adoption, because in a small firm the founder's habit is the firm's habit, and it is the cheapest, most powerful lever you have.

Helping owners get genuinely hands-on, so they can lead AI from experience rather than from a distance, is exactly what the AI Lessons for Leaders sessions are for. If you have been asking your team to do something you have not done yourself, book a personal lesson.

Sources and further reading

• Greg Shove, gregshove.com, 2025. His argument that leaders should model AI use for their teams and the people around them.
• Jacob Morgan, "What Leaders Need to Know About Where AI in the Workplace Really Stands Today," Great Leadership, 2025. A reminder that most firms are earlier in adoption than the noise suggests, and that honesty from leaders matters.