Field team dispatch automation: who goes, in what order, and when a person decides
Field team dispatch automation for service firms: match each new job to the nearest technician, share the day by area, order each round, and keep a dispatcher in the loop.
Part of the guide Workflow automation
Field team dispatch automation takes the routine part of a dispatcher's day off their desk: each new job is matched to the nearest technician or base, the next day's jobs are shared fairly by area, each technician's round is put in order, and everyone gets their list on their phone. The dispatcher stops copying addresses and starts handling the exceptions: the urgent call, the customer who needs a specialist, the van that broke down.
This article is for service businesses with a team on the road: appliance and lift repair, pest control, cleaning, security patrols, utilities, installers and home care. It covers the two kinds of dispatch (as jobs arrive, and the night before), which decisions to leave to a person, a made-up example in Kuala Lumpur, the mistakes to avoid and how an AI employee can run it.
Two kinds of dispatch
As jobs arrive. An urgent repair or a call-out comes in during the day and needs someone soon. The question is who is closest and free. Speed matters more than a perfect plan.
The night before. Planned visits, maintenance rounds and installations for tomorrow are known in the evening. The question is how to share them so nobody crosses the city twice, then which order each person visits in. A good plan matters more than speed.
Most firms need both. Automating the second one first usually saves the most time, because it is the same work every evening.
What to automate and what to keep for a person
| Step | Automate it? | Why |
|---|---|---|
| Finding the nearest base or technician to a job | Yes | Pure distance arithmetic, done the same way every time |
| Sharing tomorrow's jobs by area | Yes | Compact, even areas are hard by hand and easy for software |
| Ordering each round | Yes | Trying every order is what computers are for |
| Choosing a specialist for an unusual job | No, ask a person | Skills, certifications and customer history are rarely in the data |
| Bumping a planned visit for an emergency | No, ask a person | It is a judgement about customers and contracts |
| Telling technicians their list | Yes | The same message every morning, sent to the right phone |
A starting split. Firms with good skills data can automate more.
Two honest limits. Distance from a job to the nearest technician is worked out from fixed points you provide, such as each technician's base or home, not from live phone positions. And that distance is a straight line, so on a city scale the drive is usually longer. For the ordering of a round, the route planner does use the roads for usual traffic where the routing service is available, and marks the result as an estimate where it is not.
A worked example, made up
An appliance repair company in Kuala Lumpur has eight technicians, each starting from home, and gets about 50 jobs a day: some booked days ahead, some urgent.
- Customers book through a form on the company's website. Each booking lands with its address and how urgent it is.
- Urgent jobs are matched at once to the two nearest technicians' homes. The dispatcher gets a message with both names and picks one with a single reply.
- At 18:00, tomorrow's booked jobs are shared between the eight technicians by area, about six each, then each technician's jobs are put in order from home.
- At 18:15 each technician gets tomorrow's list on WhatsApp, in order, with the minutes between stops.
The dispatcher now spends the evening on the few jobs that need a specialist instead of drawing areas on a map. The numbers are invented to show how the pieces fit.
Mistakes that make dispatch worse
- Automating the exceptions. A rule that sends the nearest technician to a job needing a gas certification creates a wasted trip. Route those to a person.
- Ignoring where people start. Technicians who start from home need rounds from home, not from the office.
- Areas that change every night. Areas made from tomorrow's jobs shift as jobs change. If customers like the same technician, keep fixed patches in a table instead.
- No room for urgent work. A round packed to the minute leaves no space for the call-out at 11:00. Cap the stops per round below the maximum.
- Messages nobody reads. One clear list per person beats a long team email.
Dispatch as an AI employee
In xMap Flows this is two short flows that share the same tables. For jobs as they arrive:
- Form page: the booking form customers fill in, with address and urgency.
- Find nearest: the two nearest technicians' homes from a CSV of their positions, with the distance to each.
- Ask a person: the dispatcher picks one of the two by email, WhatsApp or Slack, and the run carries on with the answer.
- Email or WhatsApp: the job to the technician chosen.
For the next day's rounds:
- Schedule: every working day at 18:00.
- My data: tomorrow's jobs.
- Split by territory: the jobs shared between the named technicians by area.
- Route planner: one route for each technician, with a cap on stops.
- Traffic incidents, if you want: road closures near the jobs, added to the message.
- Email or WhatsApp: each technician's round in order.
The template Nearest ambulance to each accident shows the matching idea in its simplest form, and Road closures and storms at today's install sites adds the morning road and weather check. More on field work handled by AI employees is on the field operations page.
Questions
Does it know where my technicians are right now?
No. It works from fixed points you give it, such as each technician's home or base. Update the table if someone starts from somewhere else that day.
Is the distance to the nearest technician by road?
No, it is a straight line. For ordering each technician's round, the route planner uses the roads where the routing service is available.
How many technicians can it share jobs between?
Split by territory shares items between up to 20 named people, or a set number of reps, in compact areas of about the same size.
Can the dispatcher still change things?
Yes. Put an Ask a person or Approval step where a decision is needed, and the run waits for the answer before sending anything.