I Built a Maintenance Reminder System in a Weekend. Here's How.

I used to find out a customer was overdue for service when they called me, annoyed, asking why nobody reached out. That's backwards. The whole point of a maintenance business is that you're supposed to remember before they do. I was keeping it in a mix of a spreadsheet I updated when I remembered and a mental list of "that account's probably due soon." Neither one worked once I had more than a handful of customers. So one weekend I built a system that tracks every customer's service interval and tells me, every single day, exactly who's due and who's coming up. Here's how it actually went.

The spreadsheet wasn't lying, it just went stale

My spreadsheet had the right columns: customer name, service type, last visit date, interval. The problem was it only told the truth on the day I updated it. A week later three dates were wrong, a new customer wasn't on there at all, and I had no way to glance at it Monday morning and know who to call. I wasn't missing the information. I was missing something that checked the information for me, every day, without being asked.

I didn't want another piece of software to log into. I wanted the reminders to just show up, already written, already correct, so all I had to do was send them or tell the crew where to go.

Saturday morning: I wrote down my actual intervals

Before I touched any tool I listed out, in plain language, how often each type of customer actually needs a visit. Quarterly for one service, every six months for another, some customers on a custom schedule because of equipment age or a contract I'd written years ago. I also wrote down the exceptions: who gets a courtesy call first, who just gets a text, who always reschedules so they need extra lead time. That list, not any piece of software, was the real asset. It had been living in my head the whole time.

Saturday afternoon: I handed those rules to the tool

Then I set it up with something close to this:

"Here's my customer list with service type, last visit date, and interval for each. Every day, tell me who's due in the next two weeks and who's overdue right now. Draft the reminder message in my voice for each one, text style for most customers, a phone-call note for the ones I flagged as courtesy-call-first. Don't guess at an interval I didn't give you, and don't mark anyone due early just to fill the list."

That last part mattered as much as the rest. A reminder system that nags people early just to look busy burns trust fast. I wanted it boring and accurate: due when due, not a day before, not a day after.

Sunday: I tested it against my real book of customers

Sunday I ran my whole customer list through it and checked the output against what I already knew in my gut. It caught two overdue accounts I'd genuinely lost track of, and it correctly skipped a customer whose last visit date I'd entered wrong, which told me the system was actually checking the math and not just agreeing with me. By Sunday night it was producing a clean daily list: who's due, who's coming up, and a drafted message for each, ready for me to glance at and send.

What it actually changed

I stop finding out a customer is overdue from the customer. Every morning I get a short list instead of a feeling that I should probably check the spreadsheet sometime this week. The messages already sound like me, so sending them takes seconds instead of minutes of typing. And because nothing gets missed anymore, fewer customers quietly drift to a competitor just because nobody called them back in time. The whole setup runs on about $20 a month in tool costs, which is nothing compared to what one missed renewal used to cost me in a lost account.

This is the same pattern behind most of what I build: write down the judgment calls you're already making from memory, hand them to AI once, and let it carry the part that repeats. It matters most for any business built around a recurring visit, like HVAC companies, pool service companies, and pest control companies. Most owners I talk to have tried AI casually for a one-off email or two, but few have it actually built into how they operate day to day. That's the real difference.

See what AI can do for your industry: HVAC · pool service · pest control · landscapers · property managers