The sample memo
- To
- The owner, Canyon Ridge Plumbing & Heating (fictional)
- From
- Working Theory
- Re
- What we found, and what to do next
- Scope
- Half day on site. Office, dispatch, and a sit-down with three techs at the morning meeting.
Business snapshot
- Company
- Canyon Ridge Plumbing & Heating, a fictional residential service and repair company in West Jordan.
- People
- 18. The owner, an office manager, two dispatchers who also answer the phones, one estimator, a part-time bookkeeper, and twelve field techs.
- Work
- About 70 service calls a week, six to eight install quotes, and around 120 maintenance agreements.
- Tools
- A field service app for scheduling and invoicing, QuickBooks Online for the books, Google Workspace for email and shared sheets, and group texts for everything else.
- Where the day goes
- The office manager spends most afternoons re-keying and chasing. Techs write up jobs in the truck. The estimator keeps open quotes in his head.
The short version
- Roughly 20 hours a week go to re-keying, rewriting, and chasing. Most of it sits in two places: the QuickBooks hand-off and job write-ups.
- Only one of the five fixes is AI. Two are an integration and an automation, one is a process change the owner can make alone, and one we would leave alone for now.
- One change is small enough to be running before we leave: a daily nudge on quotes that have gone quiet.
Five opportunities, ranked
Ranked by payoff against effort. Hours are a weekly estimate for this fictional company, counted across everyone the task touches.
-
1. Completed jobs are typed into QuickBooks twice
- Effort
- Medium
- Weekly hours at stake
- About 6 a week
- Right fix
- Integration
- Do next or skip
- Do next
The situation
Every job closed in the field service app is re-entered by hand into QuickBooks Online: customer, line items, payment. That is about 70 invoices a week at four or five minutes each, plus an hour most weeks fixing the ones that do not match. The app has a QuickBooks connector. It was switched on once, made a mess of the item list, and was switched off.
What we would do
Clean up the item list on both sides so the names match, then turn the connector back on with a test batch of ten jobs before the full cutover. That is a week or two of careful work, not a rebuild. The office manager gets her afternoons back and the books stop drifting away from the schedule.
-
2. Job write-ups eat the end of every tech’s day
- Effort
- Low
- Weekly hours at stake
- About 8 a week across the crew
- Right fix
- AI
- Do next or skip
- Do next
The situation
Techs finish a call, then sit in the truck typing notes with their thumbs. The notes are short and inconsistent, so the office rewrites them before anything goes to the customer. Twelve techs, twenty to thirty minutes a day each, and the office still cleans it up.
What we would do
Let techs talk instead of type. A voice memo goes through an AI step that drafts two things: a plain summary for the customer and the internal notes in the format the office wants. The tech reads it, fixes anything wrong, and sends it. Setup is an afternoon. The rollout needs a 45-minute session with the crew and a week of checking the output before anyone trusts it.
-
3. Quotes go out and nobody follows up
- Effort
- Low
- Weekly hours at stake
- About 2 a week, plus the quotes that go quiet
- Right fix
- Automation
- Do next or skip
- Do next (the quick win)
The situation
The estimator sends 25 or so quotes a week by email and follows up on the ones he remembers. Roughly a third never get an answer either way. There is no list of open quotes. There is a spreadsheet, but it is two months behind.
What we would do
Start small: a sheet that is actually kept current, and an automatic email to the estimator every morning listing each quote older than three days with no answer. That part gets built the same day (see below). The full version, where the customer gets a polite follow-up straight from the field service app, goes into a sprint once the integration in item 1 is stable.
-
4. Friday timesheets arrive by text
- Effort
- Low
- Weekly hours at stake
- About 3 a week, most of it Friday afternoon
- Right fix
- Process
- Do next or skip
- Do next, no consultant needed
The situation
Every tech texts their hours on Friday. The office manager reconciles the texts against the job times already recorded in the field service app, then chases the three or four techs who forgot. The app has a time clock. Nobody turned it on.
What we would do
Turn on the time clock, tell the crew at Monday’s meeting that the texts stop this week, and have the office manager pull the report instead. Nothing to buy, nothing to build. We would put this in the memo and leave it with the owner. Paying us to do it would be a waste of money.
-
5. Maintenance agreement renewals live in a spreadsheet
- Effort
- High
- Weekly hours at stake
- About 1 a week, plus the renewals that slip
- Right fix
- Custom tool
- Do next or skip
- Skip for now
The situation
Around 120 agreements are tracked in a sheet the office manager checks once a month. Renewals get missed when the month is busy, and nobody knows how many lapsed last year. The owner’s instinct is to have something custom built.
What we would do
We would not build it yet. A custom tool here costs more than the hour a week it saves, and the field service app already has an agreements module that sits unused. Once the QuickBooks integration is clean (item 1), load the agreements into the app and let it send the reminders. Revisit in 90 days. If the module does not fit, a small custom tool is a fair conversation then, with real numbers in hand.
Quick win implemented today
The last stretch of every half day goes to one change small enough to finish before we leave. In this example it is item 3.
We would sit with the estimator, rebuild his quote list as a clean Google Sheet (customer, date sent, amount, status), and add a short script that emails him at 7:00 every morning with every quote older than three days that still says “sent.” Then we would send three follow-ups together, so the habit starts while we are in the room.
About an hour of work. Not impressive on its own. But it is running before we leave, the estimator owns it, and it shows the rest of the memo is a plan, not a wish list.
What we would not do
-
Build a website chatbot.
The owner asked. After-hours calls already go to an answering service, and there are maybe five a week. A chatbot that quotes the wrong price creates a dispute, not a booking.
-
Replace the field service app.
It is doing its job. The problems are the unused features and the hand-offs around it, not the app itself.
-
Roll out an AI tool to all 18 people at once.
Two techs for a week, then the crew. Adoption is the hard part, and we would rather get it right than get it fast.
-
Write a 40-page report.
This memo is the deliverable. If it takes longer to read than it took to decide, something went wrong.
Next step: a 30-minute call in two weeks to decide whether items 1 and 2 become an Implementation Sprint. Items 3 and 4 should be running by then either way. No pressure in either direction.