The self‑running
contracting business

Picture a roofing company in 2036. Follow one job through it.

Nobody's buying leads. The company's AI is watching every roof in its territory — satellite images, storm data, permit records, every roof the crews have ever touched. It knows how old every roof in the county is, what it's made of, and roughly when it's going to fail. A hailstorm rolls through, and before a competitor's door-knocker even gets off the freeway, the system has already flagged the 400 homes that got hit, texted the 60 that are customers, and started the insurance paperwork. Homeowners' own AI assistants are out collecting bids — and this company answers in seconds with a measured, priced quote with financing attached, built straight from aerial imagery.

The sale that used to drag out for weeks is done the same day. Measurements from imagery. Live material pricing. Financing pre-approved. The insurance scope worked out AI-to-AI with the carrier. The homeowner makes one clean decision and gets on with their life.

The office is basically gone. Scheduling, dispatch, permits, material orders, invoices, payment reminders, review requests — no humans touching any of it. The 30-person shop that used to carry 8 office staff now runs on one ops person handling the exceptions. Revenue per employee has tripled. The crews are the company.

And the crews stay human — longer than anything else. Nobody's building a robot that can handle a steep, wet roof anytime soon, and the trades shortage makes skilled hands worth more every year. So the whole structure flips: humans only do what humans are irreplaceable at, and AI does everything between the customer and the crew.

Here's the biggest shift: the business stops selling jobs and starts owning assets. It's not chasing storms one roof at a time anymore — it manages roofs on subscription. Monitoring, inspections, maintenance, replacements scheduled before the leak ever happens. Recurring revenue, real margins, and a company that sells for a multiple of what a truck-and-a-phone-number business gets — because it owns a data asset, not a job pipeline.

The contractor of 2036 isn't someone who does jobs. He's the infrastructure company that manages every roof in his territory — with AI running everything except the hammer.

The details change by trade — HVAC, electrical, plumbing, concrete — but it's the same layer getting automated in every single one: everything between the customer and the crew.

The labor math forces it. The skilled-trades workforce is aging out faster than it's being replaced. When craft labor is the scarcest asset in the business, every hour it spends waiting on the office is the most expensive waste there is. Businesses that maximize wrench-time per employee simply out-earn the ones that don't.

The technology is already here — it's just unevenly installed. Instant lead response, AI phone agents, automated quoting from imagery, self-scheduling, automated invoicing: none of this is speculative. It exists today. What doesn't exist yet is the contractor who has assembled all of it into one machine. That's an integration problem, not an invention problem — and integration problems always get solved.

Data compounds, and compounding is unforgiving. Every job, call, and quote a systematized company completes makes its predictions sharper and its automation smarter. A competitor who starts three years later isn't three years behind — they're behind by everything the leader's data learned in those three years, and the gap widens monthly.

The money has noticed. Private equity is rolling up the trades precisely because it believes modern systems triple the value of these businesses. The question for an independent contractor is not whether their market gets systematized — it's whether they'll own the system or sell to it.

Every industry that got a demand-and-data layer built over it tells the same story. When someone else owns the customer relationship and the data, the people doing the actual work become commodity labor working for the platform's margin — ask a taxi driver, a hotelier, a restaurant on a delivery app. In the trades, the lead brokers were the first draft of this: rented demand, owned by a middleman, sold to five contractors at once.

AI finishes that story, one way or the other. Either the contractor builds the machine — owns the data, the response layer, the customer list — or somebody else's machine owns the flow of work, and the contractor becomes a subcontractor to it.

The middle dies first: the thirty-person shop with a busy office and no system. What survives is machines that own markets — and crews that work for whoever owns the machine.

The good news: this future is buildable today, by an independent contractor, one stage at a time, with each stage paying for the next. That's the roadmap below.

One rule governs everything I build: the client owns the machine. The data, the systems, the automations, the customer list — all of it belongs to the contractor from day one. I charge for building and running the engine; the engine is yours.

This is the opposite of how the industry works today, where the dependency is the business model — cancel the lead broker and the phone stops ringing, fire the agency and they keep the ad account, the tracking, the history. Fire me and you keep everything. That's the difference between renting your growth and owning it, and it's the only honest way to build toward the future described above — because a self-running business you don't own isn't yours. You'd just work there.