The Truck Weighed 44,000 Pounds All Summer

Pull the load reports from any hauling outfit that bought dispatch software between 2005 and 2020 and you'll find a truck that weighed exactly the same thing for months. Gross weight: 44,000. Load after load, June through September, five trucks, one number.

No scale on earth is that consistent. What you're looking at isn't weights. It's the fingerprint of a required field.

The field the work couldn't fill

Picture how the number gets there, because some version of this ran at hundreds of outfits. Say the shop spreads ag lime on contract farm work, priced by the load, counted at the field edge the way that work has always been counted. Three loads on the east quarter, four on the home place. Nobody weighs anything, because nothing in the deal needs a weight.

Then the new system arrives, and the ticket screen has a gross weight field with a red asterisk on it. The form won't close without it. The driver is sitting in the cab at 6pm with a screen demanding a fact the day never produced, and he has two options: call the office and argue with a dropdown, or type a number. He types a number. By August every driver has settled on the same one, because 44,000 is roughly what a loaded tandem runs and nobody got a phone call about it.

Read that again from the driver's side, because he isn't lazy and he isn't lying. The software asserted that all hauling work produces a weight. His work produced a count. One of them had to give, and it was never going to be the work.

The old carbon book never had this problem. A paper ticket holds "3 loads ag lime, east field, spread" without complaint, because paper records what happened instead of demanding what the template expected. That flexibility is exactly what the software salesmen called the problem. It was actually the point.

What an invented number costs

A field full of 44,000s looks harmless until you follow it downstream.

Billing pulls from the system, so anything priced per ton on the same screens sits next to tonnage that's fiction. The office stops trusting the weight column, which means someone cross-checks the paper anyway, which means the system has added a step instead of removing one. Within a year the drivers are keeping the carbon books again, as the real record, and someone in the office is keying the books into the software on Fridays so the reports will fill. Two record systems, one of them false, both of them paid for.

The end comes the usual way. The office manager who sat through the configuration calls, who knows why job type has to be set to "bulk ag" before the rate will load, gives her notice in the spring. Nobody else knows the why behind any of it. The software runs out its license quietly, and the operation goes back to what worked: the books, the radio, the count at the field edge. Eighteen months, five figures in fees and setup, and the lasting artifact is a login nobody remembers.

If you bought operations software in that era, you probably have your own version of this story, and the detail that stings is that the outfit down the road, the one that never bought anything, skipped the whole episode. They kept their carbon books, lost no office manager to a configuration binder, and billed the same jobs with less drama.

The resistance was rational

So here's a claim worth arguing about: the operators who refused that generation of software weren't behind. They were pricing it correctly.

The sticker said $400 a month. The real price was rebuilding the operation around the vendor's form: drivers trained on an app they'd fight, a dispatcher retyping into screens what she used to write in the book, months of setup to teach a workflow builder about edge cases it would still choke on, and one person in the office who understood any of it, until she left. For a 12-truck outfit, that trade was bad, and the owners who smelled it and kept the paper were doing math, not nostalgia. The industry spent twenty years calling them late adopters. Late to what? The rollout that died?

What those owners were actually betting, whether they said it this way or not, was that software would eventually learn to read their paper instead of replacing it. That bet has now paid off.

The direction of conformance flipped

Operators didn't get more patient, and forms didn't get much friendlier. What changed is which side conforms. AI can read the paperwork an operation already produces, as it is: the carbon ticket with a count instead of a weight, the scale slip from the elevator, the grease-smudged book from the spreader truck. A ticket that says three loads, counted, gets recorded as three loads, counted. There's no asterisk demanding the weight the day never produced, so nobody has to invent one.

flowchart TD
  A["Work done: 3 loads of ag lime, counted, no scale"] --> B["2005-2020 app: gross weight required"]
  B --> C["Driver types 44,000 to close the screen"]
  C --> D["Invoices and reports built on a fiction"]
  A --> E["Carbon ticket: 3 loads, east field"]
  E --> F["AI reads the ticket as written"]
  F --> G["Billed by the count, the way the deal was priced"]

That flip is what Termco is built on. The tickets stay paper and the drivers change nothing: Termco digitizes the tickets as the operation already writes them, flags the ones that break a rule, holds the billing rules and the reference data, and assembles the invoice totals from what cleared. A load priced by the count bills by the count, because the rule says so, not because a form extracted a weight. And it means a shop that never digitized sits one step from current, not eighteen months behind, with no rollout in between, because nobody has to change how they work.

Open the old reports

If you ran one of these systems, the test takes ten minutes and a sort button. Export a season of tickets and sort by gross weight, or by whatever field the old form demanded hardest. Count the repeats. Every suspiciously identical value is a place the software asked for something the work didn't produce, and every one of them flowed into totals someone billed from or managed by.

And if you never digitized at all, run the cheaper test: send 10-20 of your paper tickets, straight out of the book, to hi@termco.ai, and we'll send back what we find within 48 hours. The books don't change either way. That was always the right call, and it finally counts for something.