The owner of a regional fence and deck company did not ease into our first call. “The ship is sinking and we need to evacuate,” he said. “So hopefully you can help us do that, so we don’t drown.” He was not being dramatic. He was running a real, growing business with crews in the field every day, and he had lost the ability to see whether any of it was making money.
He could even put a number on the leak. The two things he could no longer track, the hours his crews worked and the true cost of each job, were running him about a thousand dollars a day. Not in dramatic blowups. In the quiet gap between work getting done and anyone being able to see it clearly.
A graveyard of half-built systems
Here is the part most owners are too embarrassed to say out loud, so I will say it for them. He had tried to fix this before. Many times. By his own count there had been about a dozen failed attempts to get a working system in place, and more than forty thousand dollars spent on builds that never shipped.
“There’s been like a dozen iterations, failed iterations,” he told me, describing the cycle. “Let’s use Monday. Let’s use Make, let’s use Zapier. Over and over again, never works.” The Monday.com setup in particular never came together. He described that integration as a failure so complete it “almost seems malicious.” When he tallied up what he had spent chasing all of it, he landed somewhere north of forty thousand dollars. My honest answer on the call was the uncomfortable one: a hundred percent of it had ended up in the trash.
None of that was bad luck. The tools were the problem. No-code platforms and automation glue are fine until your needs outgrow them, and his had. As I told him on that call, those platforms only ever offer a shadow of what is possible once you actually get into real code. What he needed was custom software built around how his company actually works, not another link bolted onto a chain that was already breaking.
Six tools, none of them talking
When we mapped what he was really running, it was six separate systems that did not share data. Jobber for quotes and invoicing. Monday for task approvals. Zapier and Make for the automations meant to hold it all together. WorkYard for crew time in the field. QuickBooks off to the side for taxes. A single receipt might pass through three of them before it landed anywhere useful, and by then the number was stale or wrong.
That is why he could not see profit per job. The data he needed was scattered across half a dozen logins, and nothing reconciled. You cannot manage what you cannot see, and he could not see much of anything.
Quick wins first, on purpose
We did not open with a grand rebuild. The ship was sinking, so the first job was to stop the water.
The one thing that would actually end the company if it failed was payroll, so we stabilized crew time tracking and the payroll approval screen first, the piece that had to work every Sunday no matter what. We pulled every submission out of Monday.com into one clean, live sheet. When something broke, he could message us directly and it got fixed the same day. That alone stunned him. “It’s been, oh, yay, another broken system,” he said of the old way, “and hopefully it’s fixed sometime in the next six months.” By the end of the first week he told me, “This is further than we’ve ever gotten.” Trust gets built in days, not slide decks.
Then the real build: one operating system
With the bleeding stopped, we built the thing he had been trying to buy for years. A single custom application that runs the whole company, with one login.
The heart of it is live job costing. Every hour a crew works and every dollar they spend gets attached to the right job in real time, then measured against what the job was sold for. Each project shows up on one screen on a red to green gradient, so a glance tells him which jobs are healthy and which are bleeding. “That’s more functionality than I would have thought possible,” he said the first time he saw it. “That’ll make us tons of money.” He knew exactly why the old stack had never delivered it. “We never were going to be able to have this kind of functionality on Monday,” he said, “except with a Rube Goldberg machine.”
Around that core sits the rest of the operation. Set price and time-and-materials billing, with different hourly rates for different kinds of work. Partial expense allocation, so a bulk purchase used across three jobs gets split correctly instead of dumped on one. Crew timecards with receipt capture, cost coded to the job on the spot. Inventory with real cost, internal price, and customer price. Invoicing and card payments. A client portal. All of it in one place, instead of six.
Replacing Jobber
Jobber is a big, capable product, and plenty of contractors run on it. But it was one more silo he was paying for, and he found it needlessly complicated for how his company actually quotes and bills. “I’m very excited to exit Jobber,” he said. So his new system absorbed that work too. Quote, accepted job, invoice, payment, all of it now happens inside one application shaped to fit him, not the other way around. “As soon as we can get off of Jobber, the better,” he said. “We’re in one clean, beautiful system.”
The flagship: quoting by voice
Here is the feature he is proudest of, and the one that shows where this is going.
The old way to build a quote was slow. Visit the site, drive back, dig through notes, and type the whole thing out by hand. Call it two hours a quote. His early wish, half a joke at the time, was simple: “I’m just going to give quotes as I’m driving around between client meetings, just by talking.”
So we built exactly that. He talks, and an AI agent turns his words into a finished quote. He says “cedar privacy fence, a hundred and twenty feet, two gates, tear out the old one,” and the system matches each item to his own price catalog, fills in the line, prices it, and totals the quote as he speaks. It can even take the video he records walking a site and pull the scope straight out of it. A two hour job became a conversation. “I’m twice as productive right now,” he said. “We are in business.”
What actually changed
The numbers are real. Six tools became one. A thousand-dollar-a-day blind spot became a profit figure he can read at a glance. Quotes went from hours to minutes. And as I reminded him, what he has now is not something anyone else in his trade has built.
But the line that stuck with me was quieter. “This is the first time I’ve ever felt this way working with a systems person,” he said. Earlier he had put it another way: “You’re the cure. Now I want the antidote.” After a dozen failures, the relief was never really about software. It was about finally having someone steering who knew where the rocks were. “It’s going to be an unrecognizable world for this business,” he said, and he meant it as a good thing.
That is the part worth sitting with. The tools on the market are powerful, and AI is genuinely good now. But pointing a powerful tool at a struggling business and hoping is exactly what had failed him a dozen times over. The value was never the tool. It was the hand on the reins. Someone who understands how the work really runs, who catches the system when it is confidently wrong, and who carries the job all the way to done instead of to an impressive demo.
If your business has outgrown its tools, and you have a graveyard of half-built fixes to prove it, that is the kind of work we do. Not another link in the chain. The whole thing, built to fit.