Most owners are quietly making the same mistake about AI right now.
They think it’s an IT decision. So they’re waiting for their IT team — or their MSP, or “the tech guy” — to lead. To pick the tools. To tell them what to install. To take responsibility for whether it works.
That’s the category error. And it’s why most AI installations stall before they produce anything.
AI looks like a technology decision. It isn’t. It’s an operations decision dressed up as a technology decision — which means the wrong person is being asked to lead it.
Your IT team can absolutely support an AI installation. They cannot lead it.
Because leading it requires answering questions that aren’t tech questions.
What should the business produce?
Where is human time being wasted?
Which workflows are bottlenecks?
What does the output look like once the work compounds?
Those aren’t IT questions. Those are operating questions. They belong to the business leader. Not the tech team.
Most owners haven’t seen this yet. So they’re sitting on AI — paying for ChatGPT, watching their team tinker with prompts, reading newsletters — waiting for their IT team to come back with a plan. The IT team isn’t coming back with a plan. Because it isn’t their plan to make.
After fifteen years installing technology for businesses, here’s what I’ve learned about how this actually plays out:
The owners who win this don’t win because they pick better tools. They win because they realize earlier than their peers that AI isn’t a tech category. It’s an operations category. And operations categories belong to operations leaders. Which means you.
That doesn’t mean you become the AI installer. You don’t. The same way you didn’t become your own accountant when you outsourced bookkeeping. You hire installation. But you decide what gets installed, what changes operationally, what gets measured.
The IT team — yours, or ours, or anyone’s — supports the install. They are not in a position to lead it. Because leading it requires sitting in the operator’s chair.
Most owners don’t realize they’ve never stepped into that chair on AI.
They’ve been waiting for someone else to. So the work sits in the gap between the owner who hasn’t claimed it and the tech team who isn’t supposed to.
The fix isn’t a better IT team. The fix is recognizing that AI moved from the IT bucket to the operations bucket, and most of the market hasn’t moved with it.
The businesses installing AI right now aren’t the most technically advanced. They’re the ones whose owners made one specific shift — they stopped treating AI as something their tech people would figure out, and started treating it as something they themselves were going to lead.
That’s the difference between using AI and running on AI.
Using AI is what happens when nobody’s leading the decision. It looks like ChatGPT subscriptions, tinkering, scattered prompts, no business outcome. Tools without an operator.
Running on AI is what happens when someone leads. It looks like installed workflows, eliminated cost, measurable output, compounding leverage. Operations infrastructure with an operator.
The era of “using AI” is ending. The era of “running on AI” is here. Most owners haven’t noticed yet because they’re still waiting for their IT team to tell them about it.
The IT team isn’t going to tell you. Because it isn’t their decision to make.
It’s yours.
If you want to keep thinking about this with other owners working through the same questions, AI Unlocked is a free weekly conversation IMPACT runs. Details and how to join.