The symptom
Scheduling and dispatch is done by text messages and memory
The short answer
The schedule does exist — it is just stored in one person's phone and head, in a format nobody else can read. That works, sometimes for years, until that person is sick, on holiday, driving, or leaves. The risk is not inefficiency; it is that the business cannot operate for a week without one individual. That is a continuity problem wearing an efficiency problem's clothes.
The fix starts with a single shared source of truth for what is booked. One calendar or job board that the office, the crew and the customer all read from, where a booking exists as a record rather than as a message. Calendars are a solved, standardised problem — RFC 5545 has defined the interchange format since 2009 and every major system speaks it — so this step is genuinely cheap.
What comes next divides sharply. Confirmations, reminders, writing bookings into the calendar and logging arrival times are rules-based and should stop being manual. Deciding which crew takes which job, and resequencing the day when an emergency lands, is judgment built on things no system knows — who works well together, which customer will tolerate a late arrival, who is nearly out of hours. Automate the first list. Give the second list better information and leave the decision with a person.
Why texting became the dispatch system
Because it works, immediately, with no setup, and everyone already has it. That is not a small advantage, and any replacement that is slower to use than sending a text will be abandoned within a fortnight regardless of its features. The most common failure in this category is a scheduling product that is better for the office and worse for the crew, which is a product the crew will route around.
The second reason is that the texting is genuinely two things at once: dispatching the work and having a conversation about it. Software usually replaces the first and forgets the second, so the conversation moves back to text and the schedule follows it.
The record-keeping problem nobody mentions
A schedule held in text messages is also your only record of when people worked. Federal wage-and-hour rules require employers to keep specified records of hours worked, and a thread of messages on a personal phone is a poor answer if anyone ever asks for them. The relevant requirements are published in 29 CFR Part 516 and are worth ten minutes of reading.
This is often the argument that actually moves an owner, because it reframes the change from a productivity nice-to-have into removing a concrete exposure. The arrival and completion logging in the triage table below is where that gets solved, and it happens to be one of the easiest parts to build.
Triage table — yours to use, nothing to buy
Triage: what a system should own, and what it should not
Six parts of a dispatch day, scored with prioritize_automation_workflows, Soxoa's public scoring tool, on inputs describing weekly hours, predictability, judgment, error cost, data sensitivity and systems touched. The split is stark, and it is the point: the highest-value automation here is boring, and the most tempting automation is the one to leave alone.
Scored 70 or above with low judgment. Do this first; it is also the cheapest thing on the list.
The top-ranked item and the recommended first workflow. It is also the step that makes every other improvement possible.
Predictable and low-judgment, but customer-facing or record-bearing enough to keep visible.
Rules-based and high-volume. Check the outbound-contact rules before you switch it on.
Solves the record-keeping exposure as a side effect. Medium data sensitivity, so access matters.
Low predictability and high judgment. These scored 38 and below, and two of them scored in the low teens.
Looks automatable and is not, because the answers arrive as conversations with conditions attached.
Six hours a week of genuinely hard judgment across three systems. Give the dispatcher better information instead.
The lowest score on the board. This is the human's job and it should stay that way.
The tool returns these scores as a discovery aid, not a promise of savings or feasibility, and instructs you to confirm data access, exception paths, security requirements and ownership before building, then pilot on a reversible workflow with a measurable baseline. Your hours and judgment ratings will differ from the ones used here.
Terminology bridge
What consultants call this
The category you are shopping in has a name, and so does the thing you should be careful not to buy.
The software category covering scheduling, dispatch, mobile job records and invoicing for field teams.
The visual view of jobs against crews and time. What replaces the whiteboard, not the judgment.
Automatic sequencing of stops. Genuinely valuable at high stop counts, oversold at low ones.
The open standard behind calendar interoperability. It is why one shared calendar can appear in everyone's phone.
Timestamped arrival, completion and photo records. Solves the customer dispute and the record-keeping problem at once.
If you only adopt one thing from this page, adopt the shared calendar. Every other improvement on the list is built on top of it.
Common questions
Fair questions
Our crew will not use an app. Now what?
Then do not buy one that requires them to. Confirmations and reminders can still be automated from the office side, and arrival logging can run off a single tap or a text reply. Adoption by the people in the field is the constraint that decides these projects, and it is worth designing around rather than arguing with.
Can AI build the schedule for us?
It can propose one, and for a large stop count that proposal is valuable. It cannot know that a particular customer needs the experienced crew, or that two people should not be sent out together. That is why crew assignment scored 13 in the table above rather than 70.
Is automated texting to customers allowed?
There are rules, and they matter. The FTC's Telemarketing Sales Rule guidance is the plain-English starting point for automated outbound contact to consumers. Transactional appointment reminders to existing customers sit in a different position from marketing messages, and it is worth getting that distinction right before you switch anything on.
Sources
Where this comes from
- RFC 5545: iCalendar
IETF / RFC Editor
The open standard every calendar system speaks — proof that a schedule can live in one place and appear everywhere.
- Google Calendar API overview
Google Workspace developer documentation
Vendor documentation for writing jobs into a shared calendar from another system.
- 29 CFR Part 516 — Records to be kept by employers
Electronic Code of Federal Regulations
Federal recordkeeping requirements that a text-message-only schedule does not satisfy on its own.
- Complying with the Telemarketing Sales Rule
U.S. Federal Trade Commission
Read this before you automate any outbound calling or texting to consumers.
- AI Risk Management Framework (AI RMF 1.0)
National Institute of Standards and Technology
The public reference for deciding where a human checkpoint belongs in an automated process.
Last reviewed: September 1, 2026 · every source link checked on that date
Where to go next
Three useful next reads
What field-heavy businesses in the region actually automate, and in what order.
Read it →See the evidenceConstruction assessment exampleField reporting and coordination workflows assessed with explicit assumptions, ranges and a bounded pilot.
Read it →What it costsHow we workHow a bounded pilot with a written success gate is scoped before anything is built.
Read it →Free · No strings
See which part of your dispatch day is worth changing
Share your company, website, and email. We research your operations, show the sources and assumptions, and send a private assessment link when it is ready — no call required to read it.
Other things owners tell us