AI and the Remodeling Industry
AI Won't Save Your Remodeling Company. Using It on the Right 3 Numbers Will.
I am extremely bullish on AI in business. Especially contracting and home improvement.
I also watch a lot of remodeling and home improvement companies burn months of attention on it and get nothing back. Not a dollar of measurable return. Just a stack of dashboards, three half-built assistants and a lot of excited late nights.
That isn’t an AI problem. It’s an aim problem.
Find the constraint before you build anything
The biggest risk I see isn’t that a contractor “isn’t using AI yet.” It’s that they start building tools before they’ve identified the true constraint in their business.
Your constraint lives in one of five places: strategy, execution, a person, a process or capital. I promise it is not that you haven’t built your own CRM yet.
So before you open a single AI tool, ask two questions:
- What is stopping us from delivering a far better client experience?
- What is stopping us from adding profit to our bottom line?
If AI speeds up the answers to those questions, that’s a real opportunity. If it doesn’t touch them, it’s a hobby.
Model before automation
You don’t need an AI agent to automate your emails if you don’t know how to generate leads, sell those leads and produce those sales profitably.
If your business isn’t producing a net profit right now, a 0.01% technology efficiency gain is not valuable. It’s a distraction.
Here’s the order:
- Assess your business model against your actual industry’s benchmarks. Remodeling is not new construction. Home improvement is not remodeling. Know which one you are.
- Build the disciplines each department needs for that model to work.
- Define the metrics that really move the needle, and make sure they move through your supply chain: marketing creates leads, leads fuel sales, sales create production, production generates income.
- Pressure test the model with discipline, not emotion, until the metrics hit the targets.
Then go play with AI. Now you know exactly where to point it.
The three numbers I point it at
The single most valuable use of AI right now is deploying it inside your business to do one of two things. Increase revenue-generating activity by volume. Or decrease costs that don’t have economy at scale.
That boils down to three numbers.
1. Net profit
The net profit doesn’t lie. It’s the scoreboard for every decision you make as an owner, and it’s the number that tells you whether your AI work mattered.
Use AI to tighten the work that protects margin: estimating consistency, change-order documentation, pricing reviews against your target gross margin, and job-costing reconciliation that used to wait until the end of the month. None of that is glamorous. All of it shows up on the P&L.
If you can’t draw a line from an AI project to net profit, don’t build it yet.
2. Overhead that doesn’t scale
Ask the question every growing company has to answer: can this business model grow its revenue without growing its overhead indefinitely?
This is where AI earns its keep. Every hour your office spends rebuilding the same proposal, rewriting the same follow-up, retyping job notes or hunting for a document is overhead that grows with every job you add. That’s the cost AI should attack first.
One example from my own desk. I recently built a 10-day sales training system in about four hours that would’ve taken me 40-plus hours without AI. But the client service agreements, comp plans and KPI requirements across sales, marketing and production had to be decided first. By humans. AI did the grunt work. I did the thinking.
That’s what overhead reduction looks like. Same quality, a fraction of the hours, and a human owning every decision.
3. Speed to lead
Leads are fragile and expensive. Delay kills deals.
The companies that won this year didn’t necessarily have the best ads. They had the fastest response and the cleanest handoff to sales. If your competitor gets to the home first, they have a much higher chance of taking the deal.
Protect the lead like it cost you real money. Because it did.
Instant response, qualification questions, appointment confirmation, and a clean summary in your rep’s hands before he pulls into the driveway. Measure speed to lead before and after. If the number doesn’t move, the tool doesn’t stay.
Ask “should this exist?” at the speed you build
There’s an old engineering trap: optimizing a thing that shouldn’t exist. Owners turned product builders with AI are walking straight into it.
When you can build something in an afternoon, you have to ask “should this exist?” just as fast. Is it valuable beyond the prototype? Is it valuable to my business? If yes, why, how and where do I measure it?
That one question has saved me more hours than any tool I’ve built. Those hours went into the few things that should exist, and those things actually move my business.
And please, don’t stop running your business to go sell the tool you built. If it were creating the ROI you think it is, your business would be too busy for you to manage a software company on the side. Don’t let your business die to vibe-coded dashboards and assistants. Go run your business.
AI is the assistant, not the operator
AI tools aren’t magic, and they don’t make task completion instant. Not if you’re using them as a tool instead of trying to get them to think for you.
On a normal morning, 90% of my work happens inside my AI sandbox, and it still takes six hours. Create. Read. Revise. Digest. Think. Create. Review. Proof. Repeat. Complete. What took six hours would’ve taken a week two years ago. But instant? Thinking for me? No checking and reading? Never.
Don’t make your own intelligence artificial. Use artificial intelligence to highlight yours.
Expect more work, not less
Here’s the part nobody selling AI tells you. If you use AI properly in a company with a culture of performance, you’ll end up creating more work for yourself, not less.
When you speed up production, you can get more done. When you’re performance-focused, you choose to get more done. You trade the freed-up time for a faster flow of money through your supply chain. If you’re trying to grow, that should be exciting.
So train your highest performers first. Teach them how to check and balance information. Teach them to verify, edit, re-prompt and confirm accuracy before anything ships. Training is required with AI in your company, and it takes education and experience to train someone.
The scorecard
Before you start your next AI project, write down three numbers: current net profit percentage, overhead as a percentage of revenue, and median speed to lead. Measure them again in 90 days.
If they moved, keep going. If they didn’t, you didn’t build a system. You built a toy.
Systems beat trends. The net profit doesn’t lie.
