← All work
Regional lawn care company · AI enablement, ongoing

A Salesperson Doing Data Entry

Eighteen minutes per sample · ~1,800 hours removed · a year of backlog cleared

The situation

Soil samples are how this company earns its upsells. A technician collects a bag of soil, sends it to a lab, and the lab returns a PDF. That PDF is the beginning of a real conversation with a homeowner about what their lawn actually needs.

Getting from the PDF to that conversation took eighteen minutes.

Someone had to pull the numbers off the lab report by hand and enter them into a spreadsheet to work out the right chemical mix. Then into the CRM to find the lawn size. Back to the spreadsheet to calculate volume, then cost, then margin, then price. Then pick up the phone.

The someone doing this was a salesperson. Not because anyone designed it that way, but because interpreting the result required understanding it, and the people who understood it were the people who sold it.

What it was costing

The backlog had grown to roughly six thousand samples. About a year’s worth.

Six thousand samples at eighteen minutes each is 1,800 hours of manual work — around $45,000 in labor at their loaded rate, spent on copying numbers between two systems.

That’s the smaller number. Each sample that converts is worth just under $200, and their historical close rate was about 50%. So the backlog itself represented roughly $600,000 in service revenue sitting in a folder, unworked.

And the error rate was invisible. A mistyped figure meant either underselling the treatment or overpricing it. Nobody could tell which had happened, or how often.

Meanwhile a salesperson was not in the field.

What we built

A tool that reads the lab PDF, pulls the values, cross-references lawn size from their existing system, runs the chemical and pricing calculation, and drafts the customer email — in one click.

The homeowner gets a clear recommendation and a button to accept the treatment.

The salesperson no longer processes anything. They watch who has clicked and who hasn’t, and they call the people who went quiet. Which is the job they were hired to do.

The result
  • 18 minutes to one click per sample
  • A year of backlog cleared, and samples now processed on a monthly cycle
  • Manual transcription errors eliminated from the pricing path
  • ~1,800 hours of annual manual work removed
What we didn’t forecast

Two things happened that were not in the proposal.

The salesperson went back to selling and is now writing roughly $50,000 a month in new business on top of the soil sample revenue. We won’t claim all of that — some of it he might have found anyway. But he wasn’t finding it while he was in a spreadsheet.

And the training didn’t stay in sales. It reached the operations department, where the operations director built a dashboard that now organizes how the fleet and equipment are managed. Nobody asked us for that. We haven’t even quantified it yet.

That second one matters more than the first. We can forecast the hours a tool will save. We cannot forecast what happens when a capable person who has always been too busy to think finally gets the room to build the thing they’ve been describing for three years.

That’s usually the bigger number.

Constantine Consulting — Your Partner on the Ascent
Start with a conversation

What’s taking
too long at yours?

Tell us where the hours are going. We’ll tell you honestly whether there’s a build worth doing — and roughly what it would take.

Get in touch Read the other studies