Pool Route Software: Can AI Plan Routes and Stop Density?

Yes, with one correction: route software plans the routes, and AI helps you judge them. Skimmer, GorillaDesk and Jobber can all order stops to cut driving on the right plan, while ChatGPT or Claude can read an exported stop list and flag poor density. Stop density is how much paid work you fit per hour on the road.

The two jobs are easy to confuse. A routing tool takes stops, days and start points, then returns an order. It cannot tell you that Thursday is carrying four customers a long way from everyone else, or that a new enquiry would wreck an otherwise tidy route. That judgement is what a short AI analysis of your own data is for, and it usually matters more than the stop order.

Follow me on Instagram@sagnikteaches

What a routing solver does and what it leaves to you

A solver is a piece of mathematics that tries thousands of stop orders and keeps the one with the least driving. It works from addresses, a start point and an end point. Skimmer's back-office features describe exactly this: a map of all jobs, routes built for different techs through the week, and an option to optimise the driving between stops on a route, picking a start and finish such as the company office or the tech's home. Its pricing page lists that optimiser on the Scaling Up plan, not the entry plan.

Connect on LinkedInSagnik Bhattacharya

What the solver does not decide is which stops belong on which day. A weekly pool customer is tied to a day, and moving them is a customer conversation, not a calculation. So the real work splits into three layers.

Subscribe on YouTube@sagnikteaches

1. Day assignment. Which weekday each customer lives on, and which tech covers them. This sets your density, and it is mostly manual.
2. Stop order within a day. The solver's job. Fast, free once you have the software, and usually a modest gain on a route that was already sensible.
3. Live changes. Skips, repairs, a sick tech, a new customer. Software handles one-off changes and permanent ones differently, so check how yours does it before you rely on it.

The biggest savings come from layer one, not layer two. A perfectly ordered route across a spread-out area is still a spread-out route.

Stop density, defined so you can measure it

Density is not a standard figure in any app, so define it in a way you can calculate from a stop list. Three numbers are enough.

• Average minutes between stops: total drive minutes for the day divided by the number of stops.
• Drive share: drive minutes as a percentage of the whole working day, drive plus service.
• Revenue per working hour: the day's visit revenue divided by total hours from first stop to last.

Revenue per working hour is the one to manage, because it captures both the driving and the value of each stop. A route of 30 cheap stops can earn less per hour than 20 well-priced ones close together.

One Tuesday route, rebuilt with numbers

These figures are illustrative. Say a two-tech pool business has a Tuesday route for the first tech with 22 weekly stops. Each customer pays $120 a month, which is about $27.70 a visit across a typical month of 4.33 weeks. The tech spends 15 minutes on each pool.

The gain came from two small moves, not from clever software. First, the solver reordered the route and trimmed about 18 minutes of driving. Second, two Thursday customers who lived beside Tuesday's route were moved onto Tuesday, adding 30 minutes of service for almost no extra driving. Hours barely changed, the day's revenue rose by $56, and revenue per hour rose by roughly 7 percent.

Notice the sum that matters: 24 stops times $27.70 is $665, and 532 minutes is 8.87 hours, so $665 divided by 8.87 is $75. Write that calculation into a spreadsheet once and you can run it for every route in a few minutes.

What Skimmer, GorillaDesk and Jobber offer for routing

Three products cover most small pool businesses, and they price differently, which matters for a two-route company.

The pricing logic differs in a way that changes the choice. Skimmer charges per serviced location, so cost grows with customers. GorillaDesk charges per route, so cost grows with techs. Take a two-route, 140-customer firm that wants the optimiser, not only a route builder. On Skimmer that means Scaling Up: 140 locations at $2 is $280 a month. Getting Started would be $140, but its plan list shows Route Builder without the optimiser. On GorillaDesk, two routes cost $99 on Basic, $149 on Pro or $249 on Growth, and Basic's routing is listed as limited to 25 stops, so ask how that cap applies before a 24-stop Tuesday depends on it. Read both pricing pages on the day you decide, because plans and offers change.

Jobber's constraint is worth reading closely. Because it only reorders "anytime" visits, a weekly pool customer fixed to a Tuesday window may not be reordered at all. That suits a firm doing varied one-off jobs better than a purely recurring pool route. For the wider comparison of AI features in these products, see which AI features in field software are worth paying for.

One vendor is edging towards AI that routes for you. GorillaDesk's AI agents page lists a route-optimiser agent called MantisAI, but gives no description, release status or price for it beyond the per-conversation rates it quotes for its AI agents in general. Treat it as something to ask about in a demo, on your own stop list, rather than a reason to choose the product today.

Where AI genuinely helps: reading your own stop list

Export your customers with address, assigned day, tech, monthly price and service minutes. Strip names and phone numbers before you paste anything into a chat assistant, because the analysis needs only streets and figures. Then ask for density analysis rather than a route.

Below is a stop list for one pool route: street address,
weekday, tech, monthly price, minutes on site. Names removed.

Tasks:
1. Group the stops into geographic clusters of up to six.
2. Flag any stop more than 10 minutes' drive from every other stop.
3. For each weekday, estimate drive minutes and revenue per hour.
4. Suggest up to five moves of a customer to another weekday
   that would raise revenue per hour, with the reason.
State your assumptions about drive speed. Do not invent
distances you cannot calculate from the addresses.

Chat assistants are weak at exact driving distances, so treat the output as a map-reading aid, not a measurement. A realistic reply looks like the next sample (illustrative).

Cluster A (6 stops): Marlow Close, Heron Way, Fieldgate Rd
Cluster B (5 stops): Station Approach, Orchard Lane
Outlier: 14 Quarry Lane, about 20 minutes from any other stop
Move suggestion: shift 9 Heron Way from Thursday to Tuesday,
same cluster as three Tuesday stops.

What you would fix: the assistant guessed the 20 minutes. Paste the outlier into your routing software or Maps and read the true figure before you act. The value of the AI pass is the shortlist, and you verify each item.

Four more checks that catch density problems

The outlier test

List every stop and note the minutes to its nearest neighbour on the same route. Anything above roughly double your average gap is an outlier. Suppose a customer pays $110 a month and sits 18 minutes from the next stop while your average gap is 8. Out and back they add about 20 minutes to the day. At the $75 an hour from the earlier example, that is $25 of time against $25.40 of revenue a visit. They barely pay for the drive, before chemicals and equipment wear.

The new-customer detour test

When an enquiry arrives, insert the address into the day you would give them and read the added drive time. Many tools show this when you drag a stop onto a route. Compare it with the customer's revenue per visit. An extra 3 minutes is a welcome fit. An extra 14 minutes at $75 an hour is $17.50 of cost against a visit worth $27.70, which leaves very little margin once you add the 15 minutes of service itself.

The day-balance test

Compare revenue per hour across weekdays. If Monday earns $82 an hour and Friday $61, Friday is carrying scattered customers, and Monday has room. Shifting two or three Friday customers into denser Monday gaps, with a courtesy message, is usually easier than recruiting new ones.

The gate-and-access test

Some stops cost far more than 15 minutes because of locked gates, dogs or awkward access. Add an access note column and a real average minutes-on-site per customer from your tech's logs. A 25-minute stop priced like a 12-minute one will distort every density figure above.

Adding new stops without wrecking the week

A tidy rule keeps new customers from scattering your routes. Write it down and use it for every enquiry.

1. Find the nearest existing customer by address.
2. Offer that customer's weekday first.
3. Read the added drive time in your software.
4. If it is under about 5 minutes, accept at standard price.
5. If it is 5 to 12 minutes, accept with a modest travel loading or when a second stop nearby is likely.
6. If it is above 12 minutes, decline, or offer a waiting list for that area.

The thresholds are yours to set. Work out your own by dividing a customer's revenue per visit by your revenue per hour: at $27.70 and $75 an hour that is about 22 minutes of total time, service included. Take off the 15 minutes on site and you can afford a detour of only a few minutes before the visit loses money. A short spreadsheet can print the verdict. The method is the same one in scoring a new lawn customer against your route, which is worth a read if you also run a mowing round.

Two recurring-route habits compared

Take a three-person firm. Habit one: techs choose their own order each morning from memory. Habit two: the office runs the optimiser every Sunday evening for the coming week and adds new customers by the rule above. The first gives personal flexibility and no data. The second produces a record, which means you can see in a month whether average minutes between stops fell from, say, 9 to 8. A single minute saved across 70 stops a day is over an hour of paid time a day across the crew. That is the real return on routing software, and it only shows up if you measure it.

Where route plans go wrong in practice

• Bad addresses. A mistyped street sends a tech across town. Check any stop the solver places strangely before blaming the algorithm.
• Ignoring time windows. A commercial pool with a Monday-before-nine rule or a customer who needs a call ahead is a fixed constraint. Mark it.
• Over-optimising to the minute. A route with zero slack breaks at the first gate that will not open. Leave 5 to 10 percent of the day free.
• Reoptimising too often. Customers notice if their tech arrives at different times each week. Optimise the order, but keep each customer's day stable.
• Trusting a chat assistant's distances. It estimates. Use a map service or your routing tool for real figures.

Seasonal swings and one-off jobs on a weekly route

Weekly stops are the backbone, but pool work comes with extras: a green-pool clean-up, a filter replacement, an opening or closing visit, a pump swap. Each is a work order that has to land on a route without breaking it. Skimmer's back-office page says work orders can be added to a route directly from its Route Dashboard, which is the right habit: attach the job to the nearest existing day rather than creating a separate trip.

Here is a filled-in example (illustrative). A tech's Wednesday route ends in the east side of the patch. A customer two streets from the last stop reports a cloudy pool and needs a 40-minute treatment visit. Dropping it onto Wednesday after the last regular stop adds 4 minutes of driving. Booking it on Friday, when that tech is on the west side, would add about 25 minutes each way. The rule is simple: one-off work goes where the route already is, and the customer is offered that day first.

Seasonality matters for density too. In a quiet month you may have empty gaps in a day, so it makes sense to schedule one-off jobs and quotes into them. In a busy month the same gaps vanish. Check your density figures at the same point each season, otherwise a good winter number will hide a poor summer one.

A last small example is a start-of-day decision. If your tech lives to the north and the first stop is in the south, an optimised route that starts at the depot may be worse than one that starts from home. Skimmer's route optimiser lets you choose the start and finish point, so set it per tech rather than assuming the office, and ask other vendors whether theirs does the same. A 12-minute saving on the first leg, repeated five days a week, is an hour a week of unpaid driving removed.

Choosing between the three for a small firm

Start from how your work is shaped, not from the feature list. A pool-only business with weekly stops, chemical readings and service reports leans towards Skimmer or GorillaDesk because both are built around that shape. A business that also does one-off repairs, quotes and non-pool services may prefer Jobber and accept its "anytime visit" limit on route optimisation. Ask each vendor three things during a trial: how it handles a permanent change versus a one-time change, whether you can set a start and end point per tech, and how a new customer is added to an existing route. Then rebuild one real route in each trial and compare the drive minutes it reports.

For per-user and per-route cost across several products, what a small crew pays for field software sets the figures side by side. If your jobs are mostly unscheduled calls rather than weekly stops, dispatch software compared for small fleets covers that different problem, and the AI guide maps the other places AI fits a service business.

A one-hour monthly density review

Put a recurring hour in the diary on the first Monday of each month.

1. Export the previous month's stop list and actual drive minutes per route.
2. Calculate average minutes between stops, drive share and revenue per hour for each route.
3. Run the outlier test and list any customer above double your average gap.
4. Paste the anonymised list into your assistant and ask for the five best day moves.
5. Check each suggestion on a map, then choose two moves to make.
6. Write the three headline figures in a sheet so next month has something to beat.

A realistic first-month result is small: perhaps a minute off the average gap and a couple of customers moved. The compounding comes from doing it every month and from using the new-customer rule, which stops density from decaying as you grow.

Proving the new routes cut drive time

Judge the change on numbers, not on how the week felt. Compare three months of average minutes between stops, drive share and revenue per working hour before and after you adopted routing. If drive share has not fallen, the likely cause is day assignment, not stop order, and the fix is moving customers rather than re-running the optimiser. If revenue per hour improved but customers complain about changed days, you moved too many. Reverse the least valuable change and hold the day pattern steady for a quarter.

Questions pool owners ask about route software

Does pool route software use AI to build routes?

Mostly it uses a routing solver, a mathematical method that orders stops to shorten driving. That is not the same as a chat assistant. The AI part for most owners is using ChatGPT or Claude to read an exported stop list and flag clusters, outliers and uneven days, then letting the software do the actual ordering.

How many stops should a pool tech do in a day?

There is no single right number because it depends on stop type, pool size and drive time. Work it backwards from your own figures: service minutes per stop plus average minutes between stops, against the paid hours in the day. Compare routes on revenue per hour, not stop count alone.

Can I just use Google Maps to order my stops?

For a handful of stops, yes. Beyond about ten, hand-ordering misses better sequences and cannot handle recurring weekly rules, tech start points or new customers. Maps is fine for checking one awkward day; route software is better for the weekly pattern.

Should I turn away a customer who is far from my routes?

Not automatically. Price the extra drive time into the quote first. A far stop that adds 15 minutes a week costs real money, so either charge a travel loading, wait until you have a second customer nearby, or decline politely.

Further reads

• Pool Service Contract Template: Scope, Chemicals and Access Drafted With AI — Write the scope, chemicals and access terms behind each stop.
• How Can AI Help My Lawn Care Business? Five Jobs to Hand Over — Five jobs to hand over if you also run a mowing round.
• Septic Business Software: What to Ask for Pumping Logs, Reminders and AI — Another recurring-visit trade and what its software needs.
• AI for Pest Control Companies: What a Monthly Toolkit Costs — Monthly toolkit costs for a route-based service firm.
• Mobile Hairdressing Business: Diary, Travel Time and Rebooking With AI — Plan a mobile hairdresser's diary around drive time: patch days, same-street batching and rebooking texts, with two Tuesdays compared in minutes and income.
• Pool Service Business Software Compared: Skimmer, GorillaDesk and Jobber — Skimmer bills per pool, GorillaDesk per route, Jobber per user. Priced at 40, 90 and 200 pools, with the three AI phone options compared.
• Skimmer AI Phone: What It Does for a Pool Company and What It Costs — Skimmer's AI Phone priced for 40, 120 and 300 pools, with the limits the sales page leaves out and six test calls to make before launch.
• Pool Service Marketing: Spring Opening Campaigns and Reviews With AI — Past-customer opening offers, new-customer neighbourhood campaigns, service reports that lead to review requests, and the consent and opt-out rules around them.
• How to Price Pool Service With an AI-Built Pricing Sheet — A pool service pricing sheet built from volume, stop minutes, chemicals, driving and overhead, with a worked stop, seasonal margins and a margin check.
• Solar Panel Cleaning Business: Quote and Route Jobs With AI — A worked pricing sum by panel count and access class, plus where AI helps with quotes, route clustering and recurring-plan reminders.
• Bin Cleaning Business: Route Pricing and Subscription Reminders With AI — Price a bin cleaning route by stop density, structure per-bin plans, send reminders and reschedule for weather, with worked numbers and AI prompts.
• Pet Waste Removal Business: Weekly Plans, Pricing and Routes With AI — Price weekly pet waste plans from your own time per stop, then use AI for routes, gate notes and plan wording.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Skimmer back-office and pricing pages, GorillaDesk pricing, pool service and AI agents pages, Jobber pricing page and help article on route optimisation, read in October and November 2026.