Most of the talk about AI and the trades goes one of two ways. Either “robots will replace the crews,” or “nothing will change, it’s all hands-on work.” I think both are wrong, and the real change is somewhere less dramatic and more useful.
This post is my opinion, not a prediction I can prove. Nobody knows the next five years. But I’ve sat in enough contractor offices to have a clear view of where I’d place my bets.
The field stays human
Take a roofer on a steep pitch, a plumber in a crawl space, a framer squaring a wall. The work is physical, the conditions are different on every job, and the judgment calls happen with your hands and eyes on the thing. I don’t see AI swinging a hammer in a way that matters to a small contractor any time soon.
If anything, skilled trades people get more valuable in a world where more office work is automated. There’s still a house that needs to be built, and fewer people who can build it.
The office is where it lands
Think about where a remodeling or roofing company actually loses hours and money. It’s rarely on the roof. It’s in the gaps between people and software:
- A lead comes in and sits for half a day
- An estimate goes out and nobody follows up
- A change order is agreed on a porch and never written down
- Hours get re-typed so payroll comes out right
That’s reading, writing, sorting and moving information. It’s exactly the kind of work AI is already decent at, and it’s the work that keeps owners at a desk at 9pm. My bet: the office shrinks in effort, not in importance.
Three things I expect to change
1. Estimating gets a second set of eyes. I don’t expect AI to price your jobs for you. I expect it to read the scope, flag what’s missing and compare against your last similar jobs. Think of a sharp estimator looking over your shoulder, not one doing the work alone. The final number stays yours.
2. Follow-up stops being optional. Today, follow-up depends on whether someone remembered. In a few years, a contractor who doesn’t follow up on every lead and estimate will be competing against ones who always do. That’s not a technology gap. It’s a habit gap that technology closes. I wrote about the leak in Most contractors don’t have a lead problem.
3. The customer experience becomes the product. When two contractors can both build a good deck, the homeowner picks the one who answered fast, explained the plan clearly and kept them posted. Updates and communication used to cost an owner hours. That cost is dropping.
What I don’t expect
I don’t expect AI to fix a messy business. If your pricing, your scope and your handoffs are unclear, automating them just makes the mess faster. That hasn’t changed since the days of carbon copies.
I also don’t expect owners to hand over decisions that touch money or a client. That’s not how I build anything for our clients: anything that touches money or a customer waits for a person to approve it, and every run is logged. I think that stays the right way to do it.
So who wins?
My opinion: the contractors who win aren’t the most technical ones. They’re the ones whose teams are good at a few ordinary skills, because AI makes those skills worth more. Which brings me to what you can start building now.
Where to go from here
Here are five skills I’d start building on your team this year. None need new software. Take 20 minutes with your office manager or lead and score each one.
The five-skill check. Rate each from 1 (we don’t do this) to 5 (we do this every time, even when busy):
- Writing things down. Is the way you do a job (lead handling, estimates, closeout) written anywhere a new hire could follow? Automation can only copy a process that exists.
- Naming one owner. For each handoff (lead to estimate, estimate to contract, contract to job), does one named person own it? “The office” is not a name.
- Reviewing before sending. Does someone read what goes to a customer or touches money before it goes out? This is the skill that lets you use AI safely.
- Capturing data in the field. Do photos, hours and notes get recorded on the job, from a phone, the same day? Everything downstream depends on this.
- Looking at the numbers weekly. Do you look at leads, estimates out and jobs closed once a week, not once a quarter?
Add it up. Anything scored 1 or 2 is your next project, and it’s usually a people habit, not a software purchase. Fix the lowest one first, and write the date on the sheet. Check again in 90 days.
If you’d rather have a hand with the ones that need systems, like lead intake or estimate follow-up, book a free 30-minute call. We’ll show you how we’d set up that exact piece in the tools you already use, and you’ll get a written plan with a fixed price before you commit to anything. If you’re not ready, what to automate first is a good next read.