The Splinter

There is something in your hand right now that you have stopped feeling.

You have had it so long, weeks or months or maybe longer, that it has become part of the background noise. The thing you work around without thinking. The annoyance you have filed under that is just how it is.

In my experience, that thing is where everything starts.

I spent twenty years managing apartment properties. Numbers, maintenance cycles, vendor contracts, performance reviews. I was good at it. I also had a habit of collecting problems I had decided to live with. A report that took half a Tuesday to pull and was never read the way I sent it. A process that only existed in my head because there was never time to write it down. A follow-up that should have gone out two weeks ago but did not because nobody owned it.

When I finally started using AI (and I mean really using it rather than typing a test question and closing the tab)I did not start with a strategy. I did not enroll in a course or read a book about prompt engineering. I started with the thing that was annoying me that week.

That is what I want to tell you about.

The Splinter

I call it the splinter. It is the small thing that has been in your hand so long you forgot it was there. The task that takes longer than it should. The document that gets rebuilt from scratch every time because nobody ever made a template. The conversation you dread because you never quite figured out how to have it cleanly.

Every trade shop I have talked to has a version of this. The estimate that sits in the truck because writing it up takes an hour you do not have. The customer follow-up that falls through because the technician finished the job and moved to the next one and the thank-you text never went out. The service agreement that still looks like it was made in 2009 because there has never been a slow enough week to redesign it.

These are not big problems. That is exactly why they never get fixed. They are the right size to live with indefinitely.

The splinter is also the right size to solve in an afternoon with AI. And solving it is how you learn what the tool actually does.

What the First Session Looks Like

The owner I am describing—the one who has heard of ChatGPT, maybe typed something into it once, got a generic response and closed the tab—is not behind. That person has something the twenty-six-year-old who just discovered AI does not have: two decades of knowing exactly what is wrong and exactly how it should work.

The problem has never been capability. The problem has been release. There was no clean way to get twenty years of hard-won knowledge out of one person's head and into a usable format without a team, a budget, and time nobody had.

That has changed.

When that owner sits down with the right tool and describes a real problem, not a test prompt but a real problem from last Tuesday, something happens. The tool does not replace their judgment. It does not tell them how to run their shop. It takes what they already know and gives it somewhere to go.

An estimate template that used to take an hour to build from scratch takes fifteen minutes to draft, refine, and format. A service agreement that has needed an update for three years gets rebuilt in an afternoon, in plain language, with their business name on it, structured the way they actually work. The follow-up sequence that was always going to get done someday gets written out in a single session and handed off to whoever sends the texts.

None of that required a tech background. None of it required understanding how the technology works. It required knowing the problem well enough to describe it clearly, which is exactly what twenty years in the trades gives you.

The Mistake Most People Make

The most common mistake I see is starting too big.

Someone hears about AI, decides it is time to get serious, and immediately tries to automate their entire operation. They build a system before they understand the tool. They ask it to run something they cannot describe themselves. They get a generic result, decide the technology is overhyped, and go back to working the way they always have.

The splinter approach is the opposite of that. You find the one thing. The specific annoyance. The task that happens every week and takes longer than it should. You bring that to the tool. You work on exactly that, nothing else, until it is better than it was.

Then you find the next one.

The people who get the most out of AI are not the ones who implemented the biggest system. They are the ones who solved the most splinters. Each solution teaches them something about how to ask better, how to refine the output, how to recognize when the tool is helping and when they need to redirect it. That knowledge compounds.

By the time they are ready to tackle something larger, they are not starting from scratch. They are applying three months of weekly sessions.

What This Is Not

I want to be direct about something, because the AI conversation is full of noise that does not serve the people actually running businesses.

AI is not magic. It will not run your shop. It will get things wrong, and you will have to correct it the same way you would correct a new hire who is smart but does not know your operation yet. The difference is that it shows up every day, does not complain, and will work on your estimate template at ten o'clock on a Sunday night without billing you for the hour.

It is also not a replacement for experience. This is the part that gets missed entirely in most of the content written about AI for business. The tool is only as useful as the knowledge you bring to it. A person with six months in the trades and access to the best AI available will produce worse results than a twenty-year veteran with a basic tool. The experience is the asset. The AI is just the way you finally put it to work.

At its best, it lowers the cost of starting. The thing that has been sitting on the back burner because it would take three weeks to build properly now takes an afternoon to get to a working draft. That is not a transformation. That is a door being open that used to be closed.

The Question Worth Asking

If you are a trade shop owner reading this and you have not found your splinter yet, here is how to find it.

Think about last week. Not the big projects. The background stuff. The thing that took twice as long as it should have. The document you rebuilt from scratch that you have rebuilt before. The conversation that went sideways because you never quite figured out the right way to have it. The task that is always on the list and never gets done because it requires a focused block of time you never seem to have.

That is the splinter.

Start there. Not with a transformation. Not with a new system. Just the one thing. Spend an afternoon on it. See what happens.

In my experience, what happens next is hard to predict. But it almost never looks like what you expected, and it almost always changes what you think is possible.