The Filter Changes and Salt Deliveries You Keep Forgetting to Bill

The Filter Changes and Salt Deliveries That Never Make It Onto an Invoice

It is Friday afternoon and your operations lead is doing the job she does every week: sitting with a stack of truck tickets, a QuickBooks window, and the recurring-plan binder, trying to figure out who actually got serviced and who got billed for it. Two of your techs changed sediment cartridges on-site as a courtesy and never wrote a ticket. A salt drop went out Tuesday and the invoice still is not in the system. A customer with a Fleck 5600 softener installed in 2021 is three months overdue for carbon media service, and nobody flagged it, because the only place that install date lives is a paid invoice from three years ago.

You run a real business. Multiple trucks, hundreds of accounts on salt plans, whole-home systems, reverse osmosis under half the kitchen sinks in the county. The install revenue shows up because a customer signs and pays. The recurring revenue, the part that is supposed to be the annuity, leaks out in ones and twos: a cartridge changed and never billed, a salt delivery dropped and forgotten, a resin bed that ages out a year late because no report ever told you it was due. Individually these are forty and ninety dollar misses. Across six hundred accounts and twelve months, that is a five-figure hole you dig by hand every Friday.

Meanwhile the $4,200 quote you emailed Tuesday for a softener-plus-RO system is going cold, because the homeowner said she needed to talk to her husband and nobody followed up, and the Culligan rep who called her back on Thursday now has the job. This guide is about closing those leaks with software built around how water treatment dealers actually make money, and AI that does the chasing, answering, and billing you do not have hours for.

The recurring revenue your binder and your CRM (Customer Relationship Management) cannot bill

Your install base is the asset. Every softener, filter bank, and RO unit you have installed has a service clock: sediment cartridges on a six or twelve month cycle depending on water quality, carbon media that exhausts on gallons treated, resin beds and membranes with multi-year lives, salt that runs out on a predictable interval. You know all of this. What you do not have is a system that turns "installed a 1.5 cubic foot softener at the Hendersons on 3/14/2021" into "carbon service due September, generate the work order, bill it when the tech closes it."

ServiceTitan, Jobber, and Housecall Pro schedule the jobs you tell them to schedule. Their recurring features are calendar intervals: every 90 days, every 6 months, generic. None of them know that a softener on 22-grain water for a family of five hits its carbon service faster than the same unit on softer water for a couple. Legacy packages many dealers still run, like Wintac before Davisware retired it, or Smart Service bolted onto QuickBooks, store the customer and the invoice but never turn the equipment record into a due date. So the due list lives in someone's head, and heads forget.

A custom build starts with an equipment registry: every unit tied to its model, media type, install date, the water test at time of sale, and the household it serves. A rules engine computes the next service date per unit and, on the morning it is due, drops the work order into dispatch and drafts the invoice so it goes out the second the tech marks the job done. Salt deliveries bill automatically on drop confirmation from the driver's phone. Before any of this goes live, you mine the data you already have: a one-time pass over years of QuickBooks and Wintac records reconstructs your install base and hands you, on day one, several hundred accounts overdue for service you were never billing. That report alone tends to pay for the first phase.

The 9pm "my softener is flooding the garage" call

A softener sticks mid-regeneration on a Sunday night. Water is running to drain, the customer panics, and calls your main line. It rings to voicemail. She leaves a message, then calls the next dealer on Google, and by the time you check messages Monday you have lost the call and maybe the customer, who now thinks you do not answer the phone.

An answering service is not the fix. The person taking the message does not know your service area, cannot tell an emergency bypass from a routine salt reorder, and books nothing into your actual calendar. ServiceTitan has call-booking screens, but they still need a human sitting there at 9pm.

An AI phone agent answers every call, day or night. It recognizes the caller's number, pulls up their equipment, and knows the difference between "no water and it is spraying" and "I need more salt." Emergencies get triaged and booked into the real dispatch calendar in the right slot, with a tech certified for that work, and the customer gets a text confirmation before she hangs up. Routine salt reorders are handled start to finish without waking anyone. You are not paying a model to sound clever. You are paying for the phone answered at 9pm and a booked job on the board by morning.

The $4,200 system quote that sat for three days

Whole-home systems are where your margin lives, and they are exactly the quotes that stall. The homeowner wants to think about it, check with a spouse, compare the Culligan number. Your rep moves on to the next install and the estimate sits. Three days later it is dead, and you never knew whether a single follow-up would have saved it.

Your CRM stores that estimate. It does not chase it. Follow-up depends on a busy owner remembering, on the right day, to send the right message, which is why it usually does not happen.

A custom automation watches for estimates over a dollar threshold that stay open. It fires a same-day thank-you text, a two-day follow-up that references the actual system quoted and the problem it solves (the iron staining, the chlorine taste, the hardness eating the water heater), a financing nudge on day four, and a call task assigned to the rep if the quote is still open at day seven. The AI drafts each message in your dealership's voice, pulled from the quote's own line items, so it reads like your rep wrote it, not a template. The outcome you are buying is the quote out the same day and followed up without anyone having to remember.

Five-star installs nobody can find on Google

You commission forty systems a month and do clean, careful work. You also have thirty-one Google reviews, and the dealer across town has three hundred. When the next homeowner searches for a water softener, that number decides who gets the call, and it is not you.

Generic review blasts make it worse. They fire at everyone, including the customer who just called about a warranty leak, and they go out at the wrong moment, so they get ignored.

The fix is tied to the job, not a nightly batch. When a tech closes a job, retests the water, and shows the customer the before and after hardness numbers, satisfaction peaks. A review request fires then, personalized to the system installed, routed to the platform you want to grow. Sentiment gating keeps it clean: a customer who signals frustration is routed into a private service-recovery flow that pings your ops lead, not out to a public one-star. Over a few months, the review count that decides your next job starts climbing on its own.

The salt truck that drove past the same street twice

Your salt route is planned on a whiteboard and a driver's memory. The truck zigzags across town, doubles back for an account someone forgot, and burns an hour of paid drive time it did not need to. On the same board, an install gets promised to two customers in the same window because dispatch is a marker and an eraser. Your ops lead spends the first hour of every day untangling it.

WorkWave Route Manager can route deliveries, but it lives in its own world, separate from your service jobs. ServiceTitan can dispatch service, but it is not optimizing a salt route against how many bags fit on the truck and which windows customers gave you. The two halves of your day never share a map.

A custom platform puts salt deliveries, filter services, and installs on one board and one map. Routing accounts for truck salt capacity, tech certification (only some of your people set up RO), delivery windows, and drive time, and it rebalances the day automatically when a 9pm emergency lands on the schedule. The truck stops backtracking, the install stops getting double-booked, and the hour your ops lead spent untangling the board goes back into the business.

What this costs and how long it takes

These are the bands Digital Heroes sees across more than 2,000 delivered projects, not a menu price. A focused first release, the one that plugs the filter and salt billing leak and mines your existing data for the overdue list, typically runs $50,000 to $120,000 and ships in 10 to 16 weeks. A full operations platform, one that unifies dispatch and routing, the AI phone agent, estimate follow-up, reviews, and recurring billing, typically runs $150,000 to $350,000 and phases over 6 to 12 months, so you are getting value from the first module while the rest is built.

What pushes the number up in this niche specifically: modeling media life per equipment model instead of a flat interval, because your softeners and RO units do not all age the same way; salt route optimization that respects truck capacity and delivery windows, which is real logistics work; clean migration out of Wintac, Smart Service, or a decade of QuickBooks so the install history survives the move; capturing water test results and warranty coverage per component; and consolidating multiple dealer locations onto one system if you are a rollup. The more of your revenue is recurring, the faster the recovered billing pays the build back.

When ServiceTitan is enough, and when it is time to build

Be honest about where you are. If you are a smaller shop doing mostly one-off installs and straightforward service, with a light book of recurring plans and no real salt-delivery operation, ServiceTitan, Jobber, or Housecall Pro is genuinely enough, and you should not spend six figures to replace it. Those tools are good at what they were built for: scheduling a job, dispatching a tech, taking a payment.

The signals that it is time to build custom, or to layer AI on top of the CRM you already run, are specific. You have hundreds or thousands of recurring salt and filter accounts and you know you are not billing all of them. You reconstruct who got serviced from truck notes every week. Your install history is trapped in a tool being sunset or that nobody automates against. You are a private-equity-backed rollup trying to put three acquired dealers on three different systems into one view. Or your per-seat CRM fees keep climbing while the tool still cannot bill a carbon service on the right date. When that is your reality, the move is usually not to rip out ServiceTitan on day one. It is to layer a recurring-revenue engine and AI automation on top through the API, and build custom only where the tool actively blocks how you make money. Take the recovered billing from phase one and let it fund the rest.

How to choose a developer for water treatment software

Most software teams have never seen a softener. A few questions separate the ones who understand your business from the ones who will learn on your budget.

Ask them to describe, out loud, how they would schedule a carbon media service for a specific softener model on hard water. If they reach for a generic "every 90 days" recurring task, they do not understand that your equipment drives the clock. The right answer talks about equipment records, media life, and water quality, not calendar intervals.

Ask how they would get your data out of Wintac, Smart Service, or ServiceTitan and into the new system without losing install dates and equipment history. A serious team talks about a data audit before it talks about features, because migration is where these projects quietly fail.

Ask whether they have built route optimization before, with real constraints like vehicle capacity and delivery windows. Salt logistics is not a to-do list, and a team that has never done routing will underestimate it badly.

And ask, plainly, who owns the code and the data when you are done. The answer you want is that it lives in your repository and your cloud account, that you can hire anyone to work on it later, and that you are never renting access to your own customer list. If they hedge on that, keep looking.