Give the team their time back.

Two people, seven manual workflows. The largest customer's order process alone ran 25 to 30 hours a week, twice daily at 7am and 1pm. Each workflow lived partly in a document and partly on a piece of paper on somebody's desk.

The sharper version: a man hired to move, store, and track physical product was spending six to seven hours a day typing. And it was never six hours of work. It was a smaller amount of work shattered across the day by interruptions — and none of these workflows pause cleanly, so every interruption meant starting the task over. The problem wasn't that he was busy. It was that the work couldn't hold still.

I took the work over manually first, because doing it is how you find out what it actually is. The takeover was the discovery phase. You cannot automate a workflow you have only been told about.

From there: a roadmap sequenced by fastest-to-value rather than biggest-on-paper. A twice-weekly build cadence that runs the team. A weekly client update — the standing artifact that keeps scope, trust, and expansion aligned. A connective layer over their inventory system, so the team asks a question instead of logging in and digging. And the hand-maintained numbers on the paper on the desk were built into the automation, so nobody maintains that paper anymore. Work that disappeared rather than moved.

Coordination across the client, the build team, and the contracting partners was mine.

All seven workflows run end-to-end with one or two manual touchpoints each — pull a document off the printer, approve before the next step. The human keeps the judgment call; the system does the labor. That's the design, not a shortfall: handing a small business fully autonomous order submission on day one is how you lose their trust the first time it gets something wrong.

Roughly a 90% reduction in manual hours. The after-figure — three to six hours a week across all seven — is my own estimate and hasn't been audited yet.

The hiring plan had at least two roles in it, and the work was spread across enough of the team that it could have been more. Those hires didn't happen — and neither did the HR function that would have come with them.

Within a couple of weeks of the first tasks coming off, the team's hands were back on the work they were hired for.

The build itself. Diagnosis, architecture, and sequencing were mine; a build team executed. I was engaged through a contracting partner, so the client relationship isn't solely mine either.

As of September 2026 this is live and mid-flight. The after-figure needs an audit. What began as seven workflows is now an acquisition-ready operations platform, and the real claim is still being proven: that an acquired business can be turned around without hiring for it.