AI and the Remodeling Industry
Stop Buying AI Tools. Start Building AI Systems.
Every week there’s a new AI tool that promises to change your business.
And every week, some contractor buys it, logs in twice and forgets it exists.
I’m extremely bullish on AI. I’m not bullish on tools. Those are not the same thing, and confusing them is the most expensive mistake I see owners make right now.
A tool is not a system
A tool does a task. A system produces a result.
A tool writes an email. A system makes sure every lead gets a response in 60 seconds, every rep gets a brief before the appointment and every missed follow-up gets flagged before it costs you a job.
A tool sits in a browser tab. A system is wired into how your company actually works: who does what, in what order, measured by what number.
Most owners are collecting tools and calling it an AI strategy. That’s not a strategy. That’s a subscription bill.
How tool sprawl happens
It always starts the same way.
The owner sees a demo. It’s impressive. They buy it. They show it to the team. The team uses it for a week, then goes back to doing it the old way because nobody changed the process around it.
Repeat that five times and now you’re paying for five tools, your team has five more passwords and nothing on the P&L changed.
Worse, now your data lives in five more places. Which means somebody in the office is copying it between them. Congratulations. You used AI to add overhead.
Start with the process, not the product
Here’s the order I use every single time.
1. Pick one number. Net profit, overhead as a percentage of revenue, speed to lead, close rate. One. The thing that matters most in your business right now.
2. Map the process behind it. Every step, every handoff, every person. Write it down. If you can’t draw it on a whiteboard, you can’t automate it.
3. Find the constraint. Where does it slow down, break or depend on one person remembering? That’s where you build.
4. Then pick the tool. Now you know exactly what you need it to do. Most of the time, a tool you already pay for can do it.
5. Wire it in. Change the process so the tool is part of it, not next to it. Write the new SOP. Train the team. Remove the old way.
6. Measure. Before and after. 30 and 90 days. If the number didn’t move, it’s gone.
Model before automation. Process before product. Every time.
Should this exist?
With AI, you can build almost anything in an afternoon. That’s the danger.
When building gets that fast, you have to ask “should this exist?” just as fast. Is it valuable beyond the prototype? Does it move a number that matters? Will anyone actually use it next month?
If you can’t answer those three, don’t build it. The best AI system I ever built was the one I didn’t, because the process it was supposed to fix needed to be deleted, not automated.
What a real AI system looks like
Here’s a simple one. A post-sale handoff system for a remodeling company.
- The rep closes the job. The signed contract, measurements and notes land in one place.
- AI drafts the production handoff: scope, materials, customer preferences, anything promised in the home.
- The production manager reviews and approves it in five minutes instead of chasing the rep for an hour.
- The customer gets a welcome message with what happens next and when.
- The job gets flagged automatically if anything is missing before install.
No new app. No new login. Just a broken handoff fixed with tools the company already had, built around a process somebody finally wrote down.
That system protects margin on every job. Missed details at handoff are where remodeling profit goes to die.
Train people, not just prompts
The system only works if your people know how to run it.
Train your highest performers first. Teach them to check AI’s work, not trust it blindly. Teach them to verify, edit, re-prompt and confirm before anything goes to a customer.
AI doesn’t fix a team that doesn’t follow process. It just makes the gaps show up faster.
The test
Look at every AI tool you’re paying for right now. For each one, answer one question: what number did this move?
If you can answer it, keep it. If you can’t, cancel it this week.
Then pick one process, build one system and measure one number. That’s how you win with AI. Not by having the most tools. By having the fewest that actually work.
If you want a place to start, I’d point you at overhead or speed to lead. Both pay for themselves fast.
Systems beat trends.
