Prompt
You are a dispatch assistant for a plumbing business. Here are tomorrow's jobs with time windows and rough locations (neighborhoods only, no full addresses): {{jobs_list}}. My working hours: {{hours}}.

First, suggest a sensible visit order that respects each time window and groups jobs by area to cut drive time. Flag any job that cannot reasonably fit and say why — I will confirm the final route.

Second, draft four reusable customer messages using merge fields [NAME] and [ARRIVAL_WINDOW]:
(a) a day-before appointment confirmation,
(b) an "on my way" text,
(c) a "running late" text offering a new window,
(d) a "sorry we missed you" text to reschedule.

Rules:
- Do not invent addresses, prices, or job details beyond what I listed.
- Never promise an exact arrival time — always use a window.
- Keep each text under 40 words, friendly and professional.

Fill in your details and the prompt updates live — then copy.

What you get back (excerpt)

Suggested order (confirm before you commit): 1. 8-9a — Water heater, Oakwood (fixed window). 2. 9:30-11a — Faucet leak, Oakwood (same area, flexible). 3. 12-2p — Sewer camera, Riverside (fixed window). 4. 3-4:30p — Disposal install, Riverside (flexible). Flag: the 8-9a Oakwood window plus drive time makes an earlier Riverside start tight, so I grouped by area instead. On-my-way text: "Hi [NAME], Mike from North Star here — heading your way now and should arrive within [ARRIVAL_WINDOW]. See you soon."

The full workflow

  1. Paste tomorrow's jobs with time windows and neighborhoods, no full addresses.
  2. Adjust the suggested order using what you know about traffic and real job length.
  3. Save the four message templates into your texting or dispatch tool.
  4. Send confirmations the night before; fire the running-late text the moment you know.

Watch out for

Use job types and neighborhoods, not customer names, full addresses, or gate and lockbox codes, when you paste your schedule into a consumer AI tool.

AI does not know real traffic, parking, or how long a job will actually run — treat its route as a draft and never commit a customer to an exact arrival time you cannot control; always send a window.

Where this comes from

Every use case on this site is grounded in real reports from working plumbers — not invented by us.

More AI use cases for plumbers

← All 6 use cases: How Plumbers Use AI