The map

Where a trade shop leaks hours: the five places.

Not a list of tools. A map of the five places a job leaks hours between the phone ringing and the money landing, each with the symptom you can see from the office, the cause underneath it, which is never the one the symptom suggests, and a test you can run this week that puts a number on it. The number is the point. Everything else on this site is a fix for one of these five.

No email required. No gate. No statistics: the only numbers on this page are the ones your shop produces when you run the tests.

Start at leak one Put a number on yours

01

The map: a job is a chain, and the leak is between the links

Every job in a trade shop, a builder or a property management company moves through the same five places. It is won: a call, a lead, a visit, a quote. It is priced: the takeoff, the material, the labour, and the moment the accepted number becomes a job. It is run: scheduled, dispatched, done, noted. It is paid for: the invoice, the draw, the chase. And, if anyone is paying attention, it produces the next job: the handover, the renewal, the water heater that is due.

The hours do not leak inside those five places. A tech on a roof is not the leak; the office manager building a quote is not the leak. The hours leak between them, at the arrows, where what one step already knew has to get to the next step and the way it gets there is a person typing it again. Name, address, phone. The line items. The price. The tech who was there. Every arrow is a re-type, a wait, or both.

That is why the tools do not fix it on their own. A field-service product, a construction platform, a property management system, each is very good inside its own boxes and knows nothing about the arrow into the box before it. And it is why the AI does not fix it either: every AI product you will be shown is a faster front of the chain, and the leak is in the middle. The pillar on what AI can actually do for a shop like yours makes that argument in full.

So the map below is five arrows, not five features. For each one: the symptom, which is what the office notices; the cause, which is what is actually happening one step earlier; and a test, which is a count that fits in a Tuesday. Run the tests, write the five numbers down, and you have the thing most shops pay to find out.

02 · Leak 1

Win the work

Leads, the phone, the estimate, and the quote that actually goes out.

The symptom

Quotes go out three days after the visit. The office knows which ones are late and dreads the call asking. Missed calls are counted by the voicemail light, not by anyone. The owner writes the quotes at night.

The cause, one step earlier

The quote is assembled from scratch each time, from a supplier PDF and memory, in a template that does not know what the last quote said. There is no loaded price book, so the fastest quoting tool in the world would be slow. And the phone is answered by whoever is nearest, which at four on a hot day is nobody.

The test, this week

Pull the last ten quotes and write two dates beside each: the site visit and the send. Then pull the call log for one week and count the calls that rang out. Two numbers. If the gap is over a day and the rang-out count is over a handful, this is your first leak, and the fix begins with the price book, not the phone.

03 · Leak 2

Price the job

Takeoffs, material, labour, and the second entry into the job system.

The symptom

The accepted quote gets typed into the job by hand. The job note says “see quote.” The tech asks the office what was sold. Two people know the price and neither of them is the one on the roof.

The cause, one step earlier

The quote lives in one place and the job in another, and the arrow between them is a person. Usually because the quote was built outside the field-service tool, in the template, so there was nothing for the tool to convert. This is the leak every AI vendor sells past, because the demo is at the front of the chain and this is the middle.

The test, this week

Take one job from last week and follow it from the phone to the books. At each screen, write down which facts were typed for the first time and which were typed again. Count the second column. There is a worked example of exactly this, with the count and the three fixes in order.

04 · Leak 3

Run the job

Scheduling, site notes, work orders, tenant requests.

The symptom

The tech texts photos and a few words to the office, and the office types them in. The dispatcher runs the day from a sheet and opens the tool to see the board. Tenant requests arrive on whichever manager’s phone the tenant has. Nobody can say, without asking, which jobs are open past a week.

The cause, one step earlier

The field never joined the system, because the app asked the tech for six fields the office needed and gave nothing back: no history at the address, no parts on the truck, no view of tomorrow. So the tool became a clerical layer on top of the same job. On the property side, the portal was never switched on or never announced, so intake stayed on the phone.

The test, this week

Ask each tech to close their last job on the app before leaving the driveway, for one week, one crew. Ask the office to read from the tool and not the sheet for the same week. Count the texts to the office. On the property side: switch the portal on, put the link on every lease and email footer, and count requests by channel for a month.

06 · Leak 5

Keep the customer

Handover, defects, renewals, the next job from the last one.

The symptom

The maintenance agreement renews when the customer calls, if they call. The review request goes out to the customers the office remembers. The water heater installed six years ago is a record nobody looks at. The next job comes from a Google search, not from the last job.

The cause, one step earlier

The follow-up that did not happen, which is most of them. Not because anyone decided against it, but because it depends on a person remembering on the right day, and the install dates and agreement end dates were never loaded into anything that could remind them. This is the cheapest leak to close and the last one anyone closes.

The test, this week

Count the maintenance agreements or the installs older than a year in the system, then count the ones with a next-contact date on them. The difference is the leak. Then switch on the reminders your field-service tool already has, for one category, and count what comes back in a quarter.

07

What an audit adds

Five tests, five numbers, a Tuesday each. If you run them, you will know more about where your shop leaks than most owners ever find out, and for three of the five the fix is configuration you already own: the price book loaded, the techs closing on the app, the reminders and the sync switched on. Do those and stop, and this page has paid for itself.

What is hard from inside the shop is three things. Running the tests across every job type instead of the one you picked. Running them with the people who do the typing, who know where the sheet is, rather than the person who thinks they know. And putting a dollar figure beside each leak that the owner will believe, because it came from the interviews and not from a vendor’s calculator. That is the AI Audit: the map above, done for your shop, the count with real minutes and real loaded costs, and a written blueprint that ranks the five leaks by what each one recovers and says, for each, configure, connect, build, or leave it alone.

A fixed fee, $5,000, sized by headcount. If the recoverable savings in the blueprint do not cover it, you do not pay, and you keep the blueprint. The first step is a 30-minute call, and if your five numbers are small, I will say so on the call.

Book a 30-minute call → How the AI Audit works

08

Questions people actually ask

How many hours a week does a trade shop spend on admin?

Nobody can tell you that for your shop except your shop, and the published averages are for someone else’s. The useful number is the count: one job traced from the phone to the books, the re-types and the waits written down, multiplied by jobs a week. Most owners who do it for the first time are surprised in the same direction. The calculator on this site does the multiplication from four inputs.

What is the biggest time waster in a contracting business?

In the shops I map, the second entry: the same job typed into the quote, then the job system, then the invoice, then the books, by a person each time. It is invisible because each re-type is small and each one is somebody’s normal Tuesday. It is usually larger than the quoting delay or the missed calls, and it is the one no vendor is selling a fix for, because the fix is mostly configuration.

How do I reduce admin in my contracting business?

In this order: load the price book into the field-service tool you already pay for, so the quote is built there and converts to the job and the invoice; make the techs close jobs on the app before leaving the driveway; switch on the reminders and the QuickBooks sync; set the rule that the bookkeeper edits in the tool and reads in the books. None of that is software you do not own. Only then look at the phone, and only after that at anything with AI in the name.

Should I hire an office manager or buy software?

Neither, first. If the office is drowning in re-typing, a second person types twice as much and the software becomes a second place to type. Count the re-types, close the ones configuration reaches, and then see what is left. What is left is either a person’s job or a build, and the count tells you which, with a number beside it.

Which of the five leaks should I fix first?

The one with the biggest count, which is usually the second entry, and the cheapest fix, which is usually the reminders. If you can only do one thing this month, load the price book, because three of the five leaks depend on it: the quote cannot be fast without it, the job cannot inherit lines without it, and the invoice cannot be built from the job without it.

What does the AI Audit add to doing this myself?

The tests above take a Tuesday each and you can run them. What is hard from inside the shop is running them across every job type, with the people who do the typing rather than the person who thinks they know, and putting a dollar figure beside each leak that the owner will believe. That is the audit: the map, the count in interviews, and a written blueprint ranking the leaks by what each recovers. A fixed fee, $5,000, sized by headcount, not owed if the savings do not cover it.

Related