The job type is the schema: what a booking code has to carry
In short
A job type is the record that binds trade, priority eligibility, expected duration, required skills, capacity class and the questions intake must ask. Dispatch, scheduling and pricing all read from it, which is why an intake agent cannot book anything until it resolves one — and why two types sharing a name with different durations wrecks a board.
Key takeaways
- The job type is a schema, not a label. Duration, skill, capacity and priority all hang off it.
- Two types sharing a name with different durations is the most common cause of an over-booked afternoon.
- Required intake questions belong to the type, so a new type ships with its own questions or books blind.
- Split a type when its on-site time is bimodal; merge two when they never route or price differently.
- Audit with real completion times: p90 more than 2.5x p50 means the type is hiding two jobs.
A job type is the record binding trade, priority eligibility, expected on-site duration, required skills, capacity class, the questions intake must ask, and the arrival window a customer is promised. Everything downstream reads it. Until whoever took the call has resolved one, there is nothing to schedule, because none of those values exist.
Most operators discover this the slow way. The job type starts as a dropdown filled in during onboarding, grows to 60 entries nobody owns, and then dispatch stops trusting it and starts reading the notes field.
One job type, field by field
Here is a single residential HVAC diagnostic, written as the record it is rather than the label it appears to be.
| Field | Value | What reads it |
|---|---|---|
| code | HVAC-DX-RES | Every downstream record; never reuse a retired code |
| trade and segment | HVAC, residential | Skill matching, price book, reporting |
| priority eligibility | emergency, same-day, routine | Triage; the tier is chosen per call, within this set |
| default on-site duration | 90 minutes | The board, drive-time packing, capacity maths |
| p90 duration | 150 minutes | Overrun risk and whether the last slot of the day is safe |
| skill tags | refrigerant handling, residential split | Technician eligibility and apprentice pairing |
| capacity class | diagnostic | How many can be sold per technician-day, separately from minutes |
| required intake fields | unit location, age, symptom, occupancy | The agent's must-answer list before booking |
| arrival window | 2 hours | The promise made on the call and in the confirmation |
| after-hours eligible | yes, with on-call approval | Whether the overnight path may book it at all |
Read that list and the intake conversation writes itself. Read only the label 'no cooling' and every value becomes a guess made by whoever answered the phone.
Duration is the field that breaks the board
The classic failure: two job types called 'Tune-up'. One is a residential split-system visit at 45 minutes. The other is a rooftop unit on a commercial site at 150 minutes, created by a branch manager 3 years ago. Both read as 'Tune-up' in the picker. An agent books 4 into an afternoon, the technician completes 2, and the 2 customers who slip were never told.
The questions belong to the type, not to the script
Attaching required intake fields to the job type is what makes an intake agent auditable. It is not following a long script; it is filling a record and refusing to book while a required field is empty. A new job type then ships with its own questions instead of inheriting a generic list.
- Residential no-cooling asks whether the system runs at all, thermostat state, breaker state, unit age and whether anyone is home in the window.
- A third-party warranty dispatch asks for an authorisation number and covered scope before anything else, because it is a different object entirely — see why a home warranty dispatch books differently.
- Where answers change the urgency, the type lists only which tiers are permitted; the tier is decided per call using the emergency, urgent, same-day and routine definitions.
- Anything the office would otherwise phone back for is a required field. A callback to ask the unit's age costs more than the question.
Auditing a job type list you inherited
- Pull 90 days of completed jobs with real on-site minutes and compute p50 and p90 per type. Anything with a p90 above 2.5x p50 is a split candidate.
- Flag every type used fewer than 10 times in the window. Rare types are usually duplicates with different spellings, and they are what an agent picks by accident.
- Compare booked duration against actual per type. A type running consistently over its default is the one making every afternoon late.
- Check each type has an owner, a default duration, a skill tag set and a required-field list. Any type missing one of the four cannot be booked safely by an agent.
- Retire rather than delete, so historical jobs keep resolving and reporting does not silently change shape.
A structured record an automated caller must fill, with refusal when it cannot, is the containment argument we make in AI agents in production and how we scope intake work in AI agents and automation. It also fixes reporting: attribution needs the job type and the lead source field to be trustworthy, and after-hours jobs are where both go missing — see overnight bookings missing from the morning board. The cluster sits under intake and call capture, within our field service and trades work.
Frequently asked questions
Short answers to the follow-ups this page tends to raise.
What is a job type in field service software?
It is the record that defines what kind of work a booking represents, and it carries the values scheduling depends on: trade, expected duration, required skills, capacity class, permitted priorities and the intake questions that must be answered. Treating it as a display label is what causes over-booked days, because the board is packing minutes it was never given.
How many job types should a trades business have?
Fewer than most inherit — typically 15 to 30 for a single-trade operator. The test is behavioural: two types should exist only if they differ in duration, skill, capacity or price. Anything else is a reporting preference a tag can carry, and every extra picker entry is another mis-booked call under pressure.
What is the difference between a job type and a price book item?
The job type describes the visit and drives scheduling; the price book item describes what is sold and drives the invoice. One diagnostic visit can end in any of a dozen priced repairs. Collapsing them gives you either an unschedulable price list or a job type list that grows an entry every time a part changes.
What should an intake agent do when nothing matches the caller's problem?
Book a general diagnostic for the right trade, capture the symptom verbatim, and flag it for a dispatcher rather than forcing the nearest-looking type. A wrong type sends the wrong technician with the wrong parts and the wrong duration; an honest diagnostic costs one triage minute in the office.
- dispatch
- intake
- scheduling
- data model
The work behind this page
Builds from our portfolio that this page draws on.
FieldRoute
An AI field-service platform that auto-dispatches the best-matched technician, optimizes routes, and tracks first-time-fix against every SLA.
OperationsQuoteForge
An AI CPQ and proposal platform that builds enterprise quotes from your catalog, guards every discount against the margin floor, routes approvals, and generates the proposal.
Sales AIRead next
- Emergency, urgent, same-day, routine: four tiers and who defines themThere is no industry-wide definition of an emergency call. The operator writes one, and until those conditions exist as data no intake agent applies them consistently.definition
- A home warranty dispatch is not a customer callA warranty dispatch arrives with an authorisation number, a covered scope someone else defined, and a payer who is not in the house. Booking it like a retail call produces uncloseable jobs.definition
- A no-heat call got booked as a routine tune-upA job sits on the board at routine priority and the recording describes an emergency. Before retuning anything, join booked jobs to their transcripts and score them against your own tier table.diagnostic
- The calls were answered overnight and the board is empty at 7amThe overnight calls were answered and the intent captured, and the jobs still vanished in the handoff. Four counts, five ranked causes, and the report that lands before the board opens.diagnostic
Working on something in this space?
Tell us where you are in a sentence or two. We'll tell you honestly whether we're the right team, and what a sensible first slice of the work looks like.
Start the conversation