Dispatchers should start with Daily Job Alerts, a 10-minute n8n lesson that emails every tech their assigned jobs at 7am with no AI involved. It comes first because it fixes the most repetitive part of a dispatcher's morning, and it teaches the workflow basics the next three lessons build on. After that, the path adds an AI agent for follow-ups after service calls and two request forms that sort incoming jobs by urgency.
All four lessons use n8n, so this path needs accounts from the start. It is also the path where safety rules matter most: AI can sort and confirm a request, but it never tells a customer how to fix gas, electrical or water problems.
Which lesson should you take first?
Take Daily Job Alerts first. You play the office at Rapid Flow Plumbing, a made-up plumbing company, with a Today's Jobs sheet of six jobs, each with a Tech and a Tech Email, and one job not yet assigned.
You set the workflow Timezone, add a schedule trigger at 7am, read the Jobs sheet, and add a Filter so only rows where Tech is not empty pass through. Six jobs go in and five come out. A Gmail node sends each tech one plain-text email with the job, address, time and customer.
After this lesson, reassigning a job means changing one name in the sheet, and tomorrow's email goes to the right tech. No AI means nothing to second-guess.
What does the rest of the dispatcher path build?
The next three lessons add AI where reading and writing are involved: follow-up messages, then incoming requests from two kinds of trade businesses.
Step 2: How can AI write follow-ups after the day's calls?
AI Customer Follow-Up Agent (12 minutes, n8n plus the OpenAI API) builds an agent for Blue Maple Plumbing, a made-up plumbing company. You give it the day's calls in one line (Linda's water heater replacement, Tom's main drain cleaning, Maria's sump pump) and it writes a personal message for each. Then you tell it to save them to a Customer Follow-Ups sheet with Customer, Service, Follow-Up Message, Next Step and Status.
This lesson comes before the forms because it walks you through creating an n8n account and an OpenAI API key, which the form lessons assume you already have. For a dispatcher, it turns end-of-day follow-ups into a list you review instead of messages you write one by one.
Step 3: How can AI triage emergency requests?
AI Emergency Request Form (15 minutes, Lovable plus n8n) builds a service request form for Blue Maple Plumbing. Each request goes to an agent whose system message sets three priorities:
- Emergency: active leak, burst pipe, sewage backup or no water.
- Same-day: no hot water, a clogged main drain or a sump pump that won't run.
- Scheduled: everything else.
The agent saves the request to a Service Requests sheet (Name, Email, Problem, Address) and emails the customer a confirmation with the priority and next steps. This is the heart of the path for a dispatcher: the triage rules are written down once, and every request is sorted the same way before you look at it.
Step 4: Does the same idea work for a different trade?
Yes. AI Service Request Form (15 minutes, Lovable plus n8n) builds the same kind of form for Cedar Ridge Heating & Air, a made-up HVAC company, but with job types instead of priorities: Repair, Maintenance, Install or Emergency. No heat, no cooling in extreme heat, or water leaking from the system counts as an emergency. Requests land in a Leads sheet with Name, Email, Request and Phone.
Taking both form lessons shows you that the triage rules are yours to write. The workflow barely changes; the system message does.
How fast should a dispatcher go?
One lesson a week is a good pace, because dispatch can't stop while you learn. The four lessons total about 52 minutes, but leave time to run Step 1 on a copy of your real schedule before you trust it.
What you need to run these for real:
- Step 1: an n8n account and a Google account. No AI and no API key.
- Step 2: adds an OpenAI API key. The API is pay-as-you-go and billed separately from any ChatGPT plan.
- Steps 3 and 4: add a Lovable account for the form.
n8n Cloud and Lovable both have paid plans, and n8n can also be self-hosted. Check each vendor's pricing page, and get the owner's approval before setting anything up on company accounts. If you want a ChatGPT-only warm-up first, Prompt Writing 101 takes 8 minutes and needs no other accounts.
What should a dispatcher do after the path?
Run the triage form alongside your current intake for two weeks, compare its priority against yours on every request, and adjust the rules until they match how your company actually dispatches. A person should still read every request before a tech is sent.
Prompts that fit dispatch: a daily job briefing for techs, emergency request triage, a service call follow-up, a service delay update and an appointment reschedule reply. Terms worth knowing: schedule trigger, filter node, AI triage, webhook and guardrails.
Trade-specific versions live in the plumbing, HVAC and electrical hubs. For further reading, see AI for Dispatchers: Job Emails, Triage and Follow-Ups, How to Sort Requests with AI: Writing Triage Rules and Automate Customer Follow-Ups After Every Service Call.
What should a dispatcher never hand to AI?
Never let AI give a customer repair instructions for gas, electrical or flooding problems, and never let it decide on its own that a call can wait. AI sorts and confirms; a person dispatches.
- Safety wording is written by people. The lessons put fixed safety lines in the system message: the HVAC agent tells anyone who mentions a gas smell to leave the house and call the gas company from outside, and the plumbing agent tells emergency callers how to shut off their main water valve. In your own version, have the owner or a licensed tradesperson approve the exact wording, tell the agent to use it word for word, and forbid any other repair advice. Emergency Instructions in AI Customer Emails goes deeper.
- Downgrading emergencies. If a request is unclear, it should go to a person, not to Scheduled.
- Arrival promises. Don't let the agent promise a time a tech hasn't confirmed.
- Customer details. Addresses, phone numbers and gate or lockbox codes stay out of chat tools. Mask names when testing prompts, and check your plan's data controls.
Key takeaways
- Dispatchers should start with Daily Job Alerts, a 10-minute n8n lesson that emails each tech their assigned jobs at 7am without any AI.
- The path then adds a follow-up agent for completed calls and two request forms that sort jobs by priority or job type.
- Every lesson in this path uses n8n, the agent steps need a pay-as-you-go OpenAI API key, and the form steps add Lovable.
- Triage rules are plain sentences in a system message, so the dispatcher can rewrite them to match how the company really works.
- AI must never give repair instructions for gas, electrical or flooding problems; approved safety wording is written by people and copied exactly.
Frequently asked questions
- Why start with a lesson that has no AI?
Because the morning job email is the most repetitive dispatch task, and a fixed rule handles it reliably. It also teaches triggers, sheets and filters, which the AI lessons reuse.
- Will the triage form replace the dispatcher?
No. It sorts and confirms each request so you see the urgent ones first. A person still reads every request and decides who goes where.
- What if a customer describes a gas smell or sparking outlet?
The agent should send only the safety wording your company approved, mark the request as an emergency and leave everything else to a person. It should never suggest a repair.
- I don't work at a plumbing or HVAC company. Is this path still useful?
Yes. The same pattern fits any business that sends people to jobs. Swap the categories in the system message for your own, and see the electrical hub for another trade's version.