Dispatch SOP Template for Service Businesses
A field-tested dispatch SOP template for HVAC, plumbing, and electrical shops. Daily, weekly, and emergency-call procedures with 2026 benchmarks.
Pre-shift: 06:30 to 08:00
Dispatcher arrives 90 minutes before first dispatch. (1) Review yesterday's incomplete jobs and reschedule. (2) Confirm every tech availability via SMS or app. (3) Pull weather, traffic, and emergency forecast. (4) Lock the dispatch board for the day with target start times. (5) Send each tech their first 2 jobs by 07:30. The 90 minutes is mandatory; without it, the day reacts instead of plans.
In-shift: 08:00 to 17:00
Dispatcher operates the board. Standard 30-minute pulse. Check tech progress, expected ETAs, customer arrival windows. Handle inbound emergency calls within 5 minutes of receipt. Update the board on every dispatch, return-to-yard, or job completion. End of every hour: log any KPIs hit (OTA, FTF flag, capacity). Standard re-route window: 60 minutes between job creation and tech arrival for emergency calls. The board exists to protect billable time: route optimization typically cuts travel 10 to 20 percent, and every gap the dispatcher closes between jobs is utilization recovered. A tech sitting idle or driving is not earning the $150K to $250K of annual revenue a productive HVAC tech generates.
Post-shift: 17:00 to 18:00
End of day debrief. Verify every job from the day is closed in the FSM. Confirm tomorrow's schedule against capacity. Note tomorrow's known risks (weather, holidays, callouts). Send the dispatcher status report to owner and ops manager. The post-shift is what enables the pre-shift the next morning.
Emergency call protocol
For trades with emergency dispatch (HVAC summer, plumbing 24/7), define a tight protocol. (1) CSR captures the call with time-stamped detail; the response clock starts the moment the phone is answered, not when a truck rolls. (2) Dispatcher has 5 minutes to identify next-available tech in radius. (3) ETA confirmed to customer in under 10 minutes from call. (4) Tech updates with arrival; dispatcher updates customer. Most operations default to closest-first dispatch, but watch for the case where a busy tech finishing nearby beats an idle tech driving across town. Skipping any of these steps costs the next 3 hours of fluid dispatch.
KPIs the SOP enforces
Six metrics this SOP supports. On-time arrival rate (target 90 percent; the industry baseline sits near 67 percent, so this is real work). First-time fix rate (target 80 percent; 70 to 80 percent is average, 89 percent and up is best-in-class, and each avoided HVAC callback runs roughly $650 in labor, drive, and lost billable time). Daily revenue per tech (benchmark for trade). Calls dispatched per dispatcher per day (capacity check). Re-dispatches per day (friction signal). End-of-day jobs not closed in FSM (data integrity signal). Remember that FTF is a lagging indicator: it is decided upstream in scheduling, dispatch, and truck stock, not by tracking the number more often.
How agentic AI takes this further
A dashboard widget that flags "job running long" is not an agent, it is a threshold alert with better styling, and it stops exactly where the SOP above needs help most, the moment someone has to actually decide what to do about it. A canned rule that auto-texts the next customer when a job status changes is automation, a fixed trigger-then-sequence, fine for the routine case, but it has no read on whether the delay is a five-minute non-issue or the start of the day falling apart. A real dispatch agent perceives the board continuously through the pre-shift and in-shift windows described above, decides when a tech running behind or a new emergency call actually means the plan needs to change, and drafts the re-route with a new ETA, then stops, because the loop only counts as an agent if it stops somewhere for a person.
During the in-shift window, the agent reads the same signals the SOP's 30-minute pulse check asks the dispatcher to read manually: job status and GPS pings off the FSM, tech progress against expected arrival windows, and new inbound calls tagged by urgency at intake. During pre-shift, it reads yesterday's incomplete jobs and today's confirmed tech availability, the same two data points step one of the pre-shift routine asks a human to pull by hand. Reading those signals continuously instead of on a 30-minute cadence is the entire point, it catches the drift between pulses, not after them.
The judgment call mirrors the SOP's own emergency protocol: does a specific delay or new call actually require a re-route, evaluated against the shop's real board, which techs are near the affected address, what skills and truck stock they carry, not a generic assumption that closest-tech-wins. A tech running behind ahead of a flexible-window job is noise. A tech running behind ahead of a customer who was promised a tight window, or a new emergency call with no idle tech obviously available, produces a drafted re-route. Anything ambiguous, two techs equally close with an unclear stock match, produces a flag instead of a guess, because the SOP's whole design goal, protecting billable time without breaking commitments, fails the moment the agent starts guessing.
Acting means an idempotent write back to the FSM, updated tech assignment, new customer ETA, drive-time estimate, each keyed to the job ID so a retry after a dropped webhook never double-books a slot or fires a duplicate text. A small Node service normalizes the FSM's webhook payloads before n8n processes them, so a platform update that renames a field breaks one script instead of the whole pipeline. If the write fails or the FSM returns something unexpected, the agent retries with backoff and then alerts the dispatcher directly instead of leaving a half-completed change in the board, the digital equivalent of the SOP's rule that every dispatch gets logged before moving on.
The checkpoint is the pre-shift and in-shift SOP above, just automated at the first step. The agent proposes the move, pings the dispatcher on Slack or SMS through Twilio, and the dispatcher approves or edits it. Nothing moves to a tech phone without that tap, because the SOP above exists precisely to keep a human owning the board, and an agent that skips the approval step defeats the reason the SOP was written in the first place. We build these in n8n against the shop's own FSM API, with every proposed and approved change logged to Supabase so the post-shift debrief has a real audit trail instead of a dispatcher's memory of what happened. Pineido builds AI agents for small business around exactly this kind of dispatch-board workflow, source handed over, running on the dispatcher's own accounts, not a vendor's.
It will not decide to skip a job, override a dispatcher's read on a difficult account, or run the emergency-call protocol's five-minute triage without a person confirming the move, those judgment calls stay exactly where the SOP puts them. What it reliably does is hold the board steady between the SOP's scheduled pulse checks, catching the drift a dispatcher juggling a full board of techs would otherwise only notice on the next 30-minute review. Run well, that shows up as fewer re-dispatches logged as a same-day surprise and a cleaner post-shift debrief, not a specific number we would print without having run it on your board first.
Get operations insights weekly
Join service-business operators who get one actionable insight on ops, automation, and growth every week.
No spam. Unsubscribe anytime.
You’re subscribed! Look for our next issue in your inbox.
That did not go through. Please try again, or email [email protected].
Ready to ship?
Tell us where you are stuck. We will review it and send a focused recommendation within one business day.
Frequently asked questions
How big is the dispatch team this assumes?
1 dispatcher per 8 to 12 active techs. Practitioners report stretching to 15 or even 20 techs per dispatcher, but only when software handles tracking and ETAs and separate staff enter sold work and parts availability. If the dispatcher is also acting as CSR and issuing POs, stay at the low end. Past 12, add a second dispatcher with territory split. Past 24, add a dispatch supervisor.
Does this SOP work for emergency-only trades like locksmith or towing?
Partially. The pre-shift and post-shift compress; the in-shift emergency protocol becomes the entire workflow. Use the emergency call protocol section as the primary workflow.
Can I use this without a real FSM?
Spreadsheets work up to 3 to 5 techs but the SOP's data integrity checks become hard to enforce. Above 5 techs the SOP needs an FSM to support it. As of 2026, Jobber starts around $39 per month and Housecall Pro around $59 per month with published pricing; ServiceTitan and FieldEdge are quote-based and aimed at larger shops (ServiceTitan typically for $3M plus revenue). Note that route optimization is not always in the base plan, so confirm it is included before you rely on the dispatcher saving travel time.
How does Pineido help adopt this SOP?
We deploy the SOP during a 30-day dispatch optimization sprint. Includes dispatcher coaching, KPI dashboard buildout, and 14 days of live shadow.