Septic Business Software: What to Ask for Pumping Logs, Reminders and AI

Septic business software should hold a record for every tank (size, compartments, lid location, last pump date, next due date), log each pumping job with volume pumped, disposal ticket and photos, send reminders from the due date, and route trucks within their capacity. AI helps with notes, wording and data tidy-up, not with routing or regulator reports.

Most demos show a calendar and an invoice. Judge the tool on the three things a pumping business cannot fake in a spreadsheet for long: the next-due date for every tank, the load a truck is carrying, and the record you may have to hand a regulator. Bring three of your own jobs to every demo and make the vendor enter them live. A polished screen with empty fields is the commonest way to buy the wrong system.

Follow me on Instagram@sagnikteaches

Why the next-due date decides everything else

Every feature a septic firm cares about hangs off one calculated field: the date a given tank is next due. Reminders read it, the schedule fills from it, and an annual plan renews against it. Widely quoted guidance puts a typical household tank at a pump-out every three to five years, and a system with a pump, float switch or mechanical parts at an inspection about once a year. So the software needs more than one interval rule, and it needs a trustworthy record of when each tank was last done.

Connect on LinkedInSagnik Bhattacharya

Here are illustrative numbers. A two-truck firm has 4,500 tanks on its books and an average three-year cycle: 1,500 pump-outs due a year, 125 a month, about 29 a week. Two trucks doing six jobs a day, five days a week, can manage 60, so planned work fills under half the capacity and the rest goes to new customers, call-outs and inspections. Now suppose 18% of the records have no usable last-pump date, which is normal after years of paper tickets and spreadsheet imports. That is 810 tanks that never reach a due list, so they never get a reminder. If a campaign to find their dates and remind them persuaded 20% to book a $380 pump-out, that is 162 jobs and $61,560. The 20% is an assumption and not a promise, which is why you test on a batch of 100 first. What matters is where the money sits: in records the software cannot act on.

Subscribe on YouTube@sagnikteaches

So the first test of any system is what it does with a tank that has no last-pump date, and how fast you can fix 810 of them, long before you look at its calendar.

The tank record: fields to insist on

Ask to see these fields on a real record, not a screenshot. Then ask whether each one can be filtered on and used to trigger a reminder, because some custom fields only sit on the screen and do nothing else.

Three traps show up repeatedly. A tank size held as free text cannot be sorted, summed or checked against a truck's capacity. One interval per customer account, rather than per tank, breaks the moment a landlord owns a standard tank and a holding tank. And gate or lid codes dropped into a general notes box are visible to every member of staff, so ask where sensitive access details live and who can see them.

The pumping log: what a single job must capture

The log is the legal and commercial record of the visit, and it is also the raw material for the customer's service report. Check that one job can capture all of this without the technician typing paragraphs:

• Arrival and departure times, truck and driver. Time on site per tank feeds your pricing and your routing. Without it you plan from guesses.
• Volume pumped, and how it was measured. A truck meter reading, a gauge and an estimate are three different things. The log should say which, because the volume feeds the truck's load and some regulators ask for it.
• Condition on opening. Liquid level before pumping, sludge and scum depth if you measure them, inlet and outlet baffles, effluent filter and any sign of backup. Dropdowns for common findings, free text for the rest.
• Before and after photos with time and location stamps. Two at minimum per tank: lid open, and the finished job.
• Disposal ticket and destination. The load or ticket number, the site and the time, tied to the job so the paperwork matches the pumping.
• Customer sign-off. A name and a signature, or a text confirmation when nobody is home.
• Recommendations raised. A cracked lid, a missing baffle, a riser worth fitting. Each should create a quote line or a follow-up task, not just a sentence buried in a note.
• Next due date, and the reason if it differs from the default. If the technician shortens the interval, the log records why, so a reminder two years later does not look like a mistake.

Then test it offline. Put a phone in flight mode and complete a job: add a photo, enter a volume, take a signature. Crews work where coverage is poor, so the job must sync when the signal returns, with nothing lost and no duplicate. For what to do with the notes afterwards, turning septic field notes into an inspection report with AI picks up where the log ends.

Reminders that start from the due date and can take a reply

A reminder feature is only as good as the rules around it. Many systems can send a text on a date; far fewer can stop sending, batch sensibly and handle what comes back. Test each of these with a dummy customer before you believe the brochure.

Keep the wording factual: the tank is due, here is the visit window, here is how to say yes. Scare wording cannot be backed up from your own records, so leave it out. Rules on whether a reminder counts as a service message or as marketing differ from place to place, so check yours before you send to anyone who has not agreed.

Routing for a truck that fills up and has to empty

A delivery van gets lighter as the day goes on. A pumping truck gets heavier, and once its tank is full it has to drive to a disposal site before the next job. A route planner built for parcels, or for plumbers, does not know that unless the vendor has built it in. Ask in this order:

• Does the planner know each truck's capacity and add the volume pumped after every job?
• Does it insert disposal trips by itself, and does it know the opening hours of each disposal site?
• Can it keep a heavy truck off the narrow lanes and weak driveways you have flagged?
• Can it cluster jobs by area and drive time, and keep one or two slots a day open for emergencies?
• Can the office drag a job to a new slot and have the driver's phone update straight away?

An illustration of why the first question comes first. A 3,000-gallon truck (about 11,400 litres) averages 850 gallons per pump-out. After three jobs it carries 2,550 gallons, and a fourth would take it to 3,400, so it has to empty first. A planner that does not track the load sequences six jobs happily, and the driver finds out at job four, forty minutes from the disposal site. If the round trip is 50 minutes and the truck costs about $95 an hour to run, one avoidable trip a week costs 50 minutes at $95 an hour, which is $79, multiplied by 52 weeks: about $4,100 a year per truck.

A chat assistant can suggest an order for five stops from a list you paste in, which helps on a bad morning. It does not know live traffic, site opening hours or what is in the truck, so do not treat it as the planner. The scheduling side of the same problem is covered in AI job scheduling and dispatch for field service teams, and automated route planning for mobile technicians shows what a costed week looks like for a van.

Regulator reports and the records you may have to keep

Some regulators require a pumping business to file a report after each job within a set number of days. Some require the owner to receive a copy, some require both, and many set a retention period for your own copies. The rules differ from one authority to the next, so the only safe test is to get your regulator's list of required fields and see whether the vendor can produce them.

• Write down your own obligations first. Which fields must a record hold, who receives it, how soon, and how long must you keep your copy? Do this before any demo.
• Ask for the report, not the promise. Have the vendor enter a test job and produce the exact report. "We handle reporting" needs a sample of a submitted one.
• Check the formats. A PDF for the owner, a spreadsheet or file layout for the authority, and whether the software submits electronically or you upload it yourself.
• Check the audit trail. Who changed the volume or the date after the job closed, and when? A record that can be altered without a trace is a liability.
• Check retention and exit. Can you export every record, with photos, in a usable format? What happens to the records if you stop paying?
• Check who updates templates when the rules change. How quickly, and at what cost?

The same discipline applies to any recurring regulated check. Backflow testing software with reminders and reports is a good second opinion on what a compliant test record looks like in another trade. Keep AI out of the filing itself: a template that refuses to save without a required field is safer than a model that fills the gap with a plausible guess.

Where AI earns its place in a septic business

AI is a fast typist with a good ear for plain English. It is a poor surveyor, a poor router and not a filing clerk. Four jobs suit it, and each comes with something to check. Strip names, phone numbers and addresses before pasting anything into a chat assistant: none of these jobs needs them.

Technician shorthand into a plain service record

Crews write fast and cryptically. This prompt turns the notes into something the homeowner can read, without letting the assistant embellish.

Rewrite this technician's shorthand as a service record for the homeowner, under 120 words.
Use only what the notes say. Do not add causes, costs or safety claims.
Where a figure is approximate, say "approximately". List recommendations separately.
Notes: [paste]

Imagine the notes read: "1000g 2comp, pumped approx 900. 1st comp sludge ~14in, scum ok. outlet baffle missing. filter cleaned. lid2 cracked, rec replace. next 3yr." An illustrative result:

We pumped your 1,000-gallon two-compartment tank and removed approximately 900 gallons. The sludge layer in the first compartment measured approximately 14 inches and the scum layer looked normal. The outlet baffle was missing and we cleaned the effluent filter. Recommendations: replace the outlet baffle, and replace the cracked lid on the second compartment. The next pump-out is due in three years.

Two things need a human. "Lid2" became "the second compartment", but it could be the second lid on the same compartment. And "baffle missing" became "replace the outlet baffle", which may be right but is a recommendation the technician never made. Ask the crew to confirm both before the record goes out, and keep the original notes stored beside the polished version.

A data audit on 30 anonymised rows

Before importing a spreadsheet, or after inheriting one, paste about 30 rows with only tank size, system type, last-pump date and notes. Ask for a list of problems, not fixes. The same approach is covered more broadly in cleaning messy data with AI.

List every row where tank size has no unit, uses an ambiguous unit, or is missing.
Do not convert or guess. Return a table: row number, what is wrong, what to ask.

The point of "do not convert" is the gallon problem in the second row. If the assistant quietly converts "1k gal" it is guessing which gallon you meant, and a 20% error in tank volume flows into pumping time and truck load. The fix belongs in the software: store volume with a unit field and refuse a save without one.

Reminder wording for different owners

One message does not suit a family in a year-round home, the owner of a holiday cottage and a landlord with several tenants. Give the assistant the segment, the tone and the 160-character limit, and ask for three versions of each. Illustrative drafts, each under 160 characters before the placeholders are filled in:

• Year-round home: Hi [first name], your septic tank is due for pumping between [date] and [date]. Reply YES and we'll offer two days, or call [number]. [company]
• Holiday property: Hi [first name], the tank at [property] is due for pumping. Want it done before your next stay? Reply YES with your dates and we'll fit around them. [company]
• Landlord: Hi [first name], two of your tanks are due: [address 1] and [address 2]. Reply YES and we'll book both on one visit and arrange access. [company]

Check each draft against the due date in the system before it is sent. The assistant writes the sentence. The date comes from the record.

Sorting replies to a reminder

Replies arrive in every shape. A chat assistant can propose a label for each one, and the office then confirms. Illustrative sorting:

The last row is the lesson. An opt-out must be handled by the software's rules the moment it arrives. Use the assistant for the ambiguous replies, with a person pressing the button. Volumes, loads, due dates and regulator fields all come from the system; the assistant drafts and sorts, and it does not decide.

Twelve written questions for septic software vendors

Put the same questions to every vendor, in writing, so the answers can be compared later. Listen for specifics. "Yes, we do that" without a screen to show is not an answer.

Records and logs

1. Is tank volume stored as a number with a unit, and can I filter on it?
2. Can I set a different interval rule for each system type, and who can override a due date?
3. Does the job log work offline, with photos and a signature, and what happens if two people edit the same tank?

Reminders and routing

1. Can reminders be released in weekly batches that match open capacity?
2. Does the planner track truck load, insert disposal trips and respect disposal site hours?
3. Where does a customer's reply land, and who sees it?

Reporting and leaving

1. Can you produce my regulator's report from a test job today, and show me one that has been submitted?
2. Can I export every record, photo and audit entry in a standard format, and is there a charge?
3. Who updates templates when reporting rules change?

AI features

1. Which tasks use AI, and which company processes my customers' data?
2. Is my data used to train models, and can I switch that off?
3. Does a person approve every AI-written message before it goes out, and is there an extra charge per use?

Scoring two systems on the same three jobs

Write the demo script before you book any demo, and give every vendor the same three jobs to enter live while you watch the clock:

1. A standard pump-out for a year-round home: tank record, log, two photos, a signature and the next due date.
2. A holding tank and a pumped system with yearly checks, to see whether different interval rules exist.
3. A landlord with two properties, a managing agent as contact and one tank already on a service plan. Then ask for next month's reminder list and see who appears on it.

Score each system 0 (cannot), 1 (can, with a workaround) or 2 (does it on screen), weight each line by how much it matters, and multiply. These two invented vendors show how the sum works. The maximum is 34.

Vendor A edges it, 27 to 26, but scored zero on a weight-three line. Set one rule before you start: any zero on a weight-three line rules a vendor out unless the fix is promised in writing with a date. On that rule Vendor B wins, and the two-point gap that tempted you to buy A disappears. Notice also how little the AI line weighs. A system with excellent AI and a broken due-date engine is a bad purchase, so switch the AI features on last, once the records, reminders and reports are right.

Two more checks cost almost nothing. Ask each vendor for a reference customer of your size who files the same regulator report, and ring them. And let one technician use the driver app for a full day before you sign: the people who will live in it should say whether the job log is quicker than their paper ticket. When the due dates are trustworthy, the marketing side gets simpler, and septic service agreements, with annual contract wording and renewals is the natural next step for turning a record into a recurring customer.

Further reads

• Septic Company Marketing: Seasonal Offers and Local Pages With AI — Use the due dates you now trust to plan offers by month.
• AI Route Planning for Small Delivery Fleets and Mobile Teams — What route tools do well, and where their limits sit.
• HVAC Dispatch Software Compared: Which Has AI Routing for Small Fleets? — A comparison of dispatch tools for small trade fleets.
• Jobber, Housecall Pro, and ServiceTitan AI Features Compared — How the general field-service apps handle AI features.
• Can AI Handle Pest Control Route Optimisation for Quarterly Customers? — Recurring-visit routing, a close cousin of pump-out rounds.
• Pool Route Software: Can AI Plan Routes and Stop Density? — What pool route software really optimises, how to measure stop density yourself, and where Skimmer, GorillaDesk and Jobber differ.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: widely quoted septic maintenance guidance from regulators on inspection and pump-out intervals. The fields, rules and figures in the examples are illustrative.