What the work actually looks like.

Since September 2025 we’ve been working inside a California Central Valley agricultural-services business that grew by combining several operations. The information had grown the way most operations’ has: a mountain of spreadsheets beside an ERP, more than one agronomy platform, and soil-moisture systems that didn’t speak to any of it.

None of those categories of system was wrong on its own. Each one was doing the job it was bought to do. The problem was that the business couldn’t see itself in one place: plans in one system, applications in another, invoices in a third, and the true answer living in whoever’s head held the newest spreadsheet.

Consolidating that stack onto a single platform, integrating the combined businesses into one operating model, and helping the operation move from selling a commodity toward selling a documented, provable service.

The order is the part worth reporting. The definitions came before the platform: what a block is called, what the word applied means here, and which system owns the plan of record once a change gets called in from the field. Every one of those was a decision somebody had to make out loud and write down, and not one of them was a software problem.

The verification habits stayed, all of them. The phone call to the field before a load ships. The weight tickets. The manual override. That’s not a workaround waiting to be designed out, that’s the culture, and an operation that trades it away for the promise of a hands-off system has given up something it can’t buy back. The software just makes the math faster to check.

The pattern here is not new, and it is not particular to agriculture. The same one showed up in oil and gas, in utilities, and in crop inputs, decades apart, each time looking like a technology problem and each time turning out not to be. The record behind that read is on the About page.

Most of what mattered above happened before anyone logged into anything.