Every owner we talk to has been pitched something with AI in the name. Very few have been told what it would take for it to work in their shop. So here is the plain version, the one you'd get across the counter from someone who would rather tell you "not yet" than sell you the wrong thing.
Readiness has nothing to do with the software
AI runs on the same three things every tool in your shop has ever run on: information it can get to, a process it fits into, and people who will use it. That's it. If your job information lives in three places and one person's head, if the way work moves through the shop has never been written down, and if the last tool you bought is quietly gathering dust, AI won't fix that. It will sit on top of it. The pitch skips this part because the pitch is about the software. Readiness is about the business.
Three questions that tell you where you stand
1. Where does the information live?
Can someone in the office say where a given job stands right now without calling the field? If the answer lives in texts, a whiteboard and somebody's memory, that's the first thing to fix, and it isn't an AI fix. An AI tool that can't reach your job information is a very polite stranger.
2. Is the process the same twice?
Quoting. Job setup. Customer updates. If how those happen depends on who's doing them that day, a tool will do exactly one thing: automate the inconsistency. Written steps come first. They don't have to be long. They have to be true.
3. Would your people use it?
The honest test is the last tool you bought. Did the crew use it? If not, why not? The usual answers are that it added work instead of removing it, that it wasn't theirs, and that training happened once, on a Friday. A process or a tool only works if people use it, and want to. AI is not exempt.
The all-or-nothing version is the only version anyone talks about. There's a sensible middle, and it's smaller than you think.
Where AI quietly earns its keep in a shop
The uses that hold up are small, bounded, and checked by a person. They take weight off your people without asking anyone to trust a machine with a decision.
- Reading and sorting the inbox nobody can keep up with, and flagging what actually needs you.
- Drafting customer updates from job notes, in your words, for a person to send.
- First-pass answers to the questions customers ask every week, with a human on the send button.
- Turning a site photo and a few words into a clean note for the file.
- Summarizing a long thread before you call someone back.
- Noticing the quote that went out and never got an answer.
Notice what these have in common. Each one replaces retyping, sorting or chasing, not judgment. Each one is easy to check. And each one gets better when your information already lives in one place, which is the readiness part again.
Where it doesn't
Pricing a job you've never done. Deciding who to hire. Talking to a general contractor you've known for twenty years. Anything where the answer has to be right the first time and nobody is checking. The version where the whole business runs on AI is sold hard, and the people selling it do not run shops. You don't need it, and you don't need to apologize for not wanting it.
The sensible middle
Here is what a sensible first step looks like for a company your size. A readiness read: what you already pay for, what's used, what's not, how information moves, and where your people spend hours a machine should spend. Then a short list, two or three places where technology or AI would amplify what you already do well or take real weight off your people, in order, with what each would take. Then a first step sized to the business. Sometimes that's a tool you already own. Sometimes it's a small custom build. Sometimes it's nothing yet. If that's the honest answer, that's the answer you should get.
What to do before you buy anything
- Pick one place where job information lives, and make it the only place.
- Write down the two or three processes that matter most, as they actually happen, not as you wish they did.
- Ask your people what eats their hours. They know, and they'll tell you if you ask plainly.
- Run the last-tool test. If it failed, find out why before the next one.
- Keep your accounting and industry software where it is. If it works, it stays.
You don't have to know what you need to start this. Most owners who call us don't. They know the business should be getting more from technology, and they know they don't want to turn a good shop into a tech company. That's enough to begin.