AI tools for a gutter company heading into fall

Every gutter company lives the same October. The first cold snap knocks the leaves loose, the phone goes from quiet to unanswerable in about ten days, and the calendar fills six weeks deep while the voicemail fills with people you'll never call back. Then it ends. By mid-December the phone is quiet again and you're wondering how much of that flood you missed.

The AI tools a gutter cleaning business needs aren't about generating demand. Fall generates the demand for free. They're about capturing a surge you can see coming from months away, and most of the capture happens before the first leaf drops. This post walks through what we'd build: a fall booking engine that fills October from your own customer list in September, a guard-sales tracker that turns every cleaning into a pitch backed by photos, and a capture layer for the storm calls that come in while both crews are on ladders.

The fall board at a glance: last year's customers queued and texted before leaf drop, October routed street by street, guard candidates flagged by tree cover, and a follow-up with debris photos going out behind every cleaning. Download as PDF

View interactive version

Sell October in September

If you're waiting for the phone to ring in October, you've already given away the best part of your season. The homeowners who paid you to clean their gutters last November are the easiest bookings you will ever get. They know you, their gutters are due, and their leaves are about to fall on schedule. Almost none of them will think about it until water is sheeting over the front porch, at which point plenty of them will panic-google and hire whoever answers.

So the machine's first job is simple. In early September, before anyone is thinking about gutters, it works through last fall's customer list and sends each one a text that reads like your office manager wrote it:

"Hi Mark, it's Ridgeline Gutter Co. We cleaned your gutters last October and with the maples over your roofline you're due again once the leaves come down. We'll be routing your neighborhood the week of Oct 19. Want your spot back?"

The reply handling is the same AI layer we use for lead follow-up everywhere else. "How much?" gets last year's price, or the new one. "Can you do the week after?" gets the next route pass through their area. "We put guards on, actually" updates the record and stops the cleaning reminders. A booked reply goes straight on the calendar in the right route block, and nobody in your office touched any of it.

Run that in September and something good happens to October: a big slice of it is already sold before the surge starts, to customers who didn't have to be won, at full price, arranged street by street instead of scattered across the county. The panic calls then fill the gaps instead of building the whole schedule. We wrote about the general version of this in the customer reactivation post; gutters are the rare trade where the comeback date is set by the trees, which makes the whole thing almost unfairly predictable.

The board that runs the season

Behind the texts sits a tracker that knows things about each house that your crew currently keeps in their heads. Stories, linear feet, tree coverage, whether it needs the tall ladder, what came out of the gutters last time. Those details set the cadence. The house under two mature oaks needs a cleaning in November and again in the spring. The newer build with one ornamental pear out front is fine once a year, and texting them twice just teaches them to ignore you.

The board shows the fall the way a dispatcher wants to see it. A due queue sorted by neighborhood so crews spend the season driving less and cleaning more. A booking bar tracking this October against last October, so by the first week of the month you know whether you're ahead or behind while there's still time to do something about it. And a flag on every house where the interval doesn't match what the crew found, because the customer who "only needs it once a year" whose downspouts were packed solid in March should quietly move to the twice-a-year list.

None of this is exotic. It's the same due-date logic we built for septic pump-out reminders, compressed from a four-year cycle down to a six-month one and pointed at a season instead of a date.

The guard pitch your crew keeps forgetting

Guard installation is where the real money in this trade sits. A cleaning is a $150 to $250 ticket. A guard job on the same house can run into the thousands, and it's sold to someone who has already hired you. And at most gutter companies the pitch happens only when a crew member remembers to make it, which on a cold roof in November with four more houses on the route is close to never.

The machine makes the pitch automatic and, more to the point, makes it land with evidence. The crew photographs what came out of the gutters. That was already worth doing for proof of service. Now, an hour after the job closes, the AI sends the homeowner those photos with a message keyed to their property record:

"Here's what we pulled out today. With that much oak canopy you'll fill up again before spring. Guards would take you out of the cleaning cycle entirely, and we can price your house from today's measurements. Want the number?"

A homeowner looking at a photo of two buckets of their own rotting leaves is the warmest guard lead that exists. The AI handles the price questions and the scheduling back-and-forth, and the tracker logs every pitch, so you can finally see the number most gutter companies have never once measured: how many cleaning customers have never been offered guards. The first time you filter that list it tends to be uncomfortable. Clean the same 120 houses every year for two years and it's common to find half were never pitched. Sixty heavy-canopy houses at guard money is six figures of pipeline sitting in your own job history.

One rule we'd wire in: the system recommends guards only where tree cover justifies them. Pitching guards to the one-pear-tree house burns trust to make a quota. The tracker knows which houses are which, so it doesn't have to guess, and neither do you.

Storm calls come in while you're on a ladder

Fall isn't only leaves. It's the season of the first big wind-and-rain storm, which yanks downspouts loose and sends water pouring over the top of clogged troughs in front of homeowners watching from the window. Those calls are worth real money, and they arrive precisely when nobody at your company can answer a phone because everyone is thirty feet up.

This is the cleanest case we know for missed call text back. The missed call gets a text inside a minute. The AI asks what's going on and where, sorts the pulled-loose downspout from the routine cleaning request, offers the next storm-repair slot for the real damage, and books the rest into the regular route. About 30% of home-service calls in this category come in after hours anyway, and in a three-week surge, business hours barely mean anything. The company whose number always responds ends up owning the street's storm work, and every storm repair becomes next year's recurring cleaning customer.

An October week with the machine on

Monday morning, the board shows 82% of the month booked, nine points ahead of last year, with route blocks colored in through the first week of November. The September send did most of that. A crew closes out a job on Sycamore Drive at 11:40, the debris photos go out at 12:45, and by 2:00 the homeowner has asked for guard pricing and been quoted from the linear-foot number already on file. Wednesday night a storm rolls through. Eleven calls come in after 7 PM. Nobody answers any of them, the text-back layer catches all eleven, three are storm repairs booked for Thursday, seven are cleanings dropped into open route slots, one is a wrong number. Thursday your office manager, whose fall used to be a headset and a legal pad, spends the morning on the two guard installs instead.

That's the whole pitch, honestly. Nothing in it is dramatic. It's a season's worth of small captures that used to leak.

When you don't need any of this

A one-crew operation cleaning the same 60 houses every fall doesn't need a dashboard. You need an afternoon in early September, your phone, and last year's invoices, and you'll get most of the reactivation win by hand. That's a real answer, and we'd rather give it to you than sell you a dashboard you don't need yet.

The machine starts paying for itself when the list outgrows anyone's memory, when you're running two or more crews whose routes need packing, or when you look at your job history and realize the guard pitch has been happening by luck. And there's a scheduling note worth taking seriously: gutter work is slow in summer, which makes July and August the right time to build this, import the history, and set the intervals. The system should be sending its first September texts the same week your competitors are waiting for the phone to ring. The full picture of what we'd wire together is on our AI for gutter services page, and the plumbing underneath is the same custom automation stack we build for every trade with a season.

The leaves keep the schedule

Gutter cleaning might be the only trade whose customers are put on a calendar by the trees themselves. The leaves will let go the same three weeks they always do. The only question is whether your October is built in September from a list you already own, or assembled one panic call at a time while the voicemail fills with jobs you'll never see.

If you want the version wired to your own customer list, tell us about your gutter business and we'll sketch it before the season turns.