Agents in the Trades
Communication

An iMessage agent that takes the job while you're on one

It's 8:40 p.m. and the owner of a three-truck plumbing shop is under a house swapping out a water heater when a homeowner three towns over, standing in a cold kitchen, texts the number off the truck: no heat. A branded iMessage agent answers in seconds, qualifies the job, books the call, and takes the deposit — while he is still under the house.

Source: Bland (@usebland on X), "iMessage agents for your business" post, Oct 15, 2026 — https://x.com/usebland

The problem

A small trades shop does not lose work in the field. It loses work at intake, while everyone is already on a job.

The owner is on a roof, in an attic, or under a house. The office is him, or one dispatcher doing everything else. A homeowner with a burst pipe, a dead furnace, or a tripping panel does not leave a voicemail and wait. She texts the number on the truck, hears nothing for two hours, and calls the next company.

Qualifying takes phone tag even on a good day: what is wrong, where, how bad, can you send a photo, when can someone come. Each gap lets the customer drift, and after hours the shop that answers first usually gets the call. The work is not hard — just timed when nobody in a small shop can do it.

How the agent workflow works

The customer texts the business number — iMessage on an iPhone — and a branded agent for the business answers in seconds, in the business's name.

It qualifies the job the way a good dispatcher would. What is wrong. How urgent — is water still running, the house without heat, anything smoking, sparking, or smelling of gas. Where the job is, and whether it sits inside the service area. It asks for photos or a short video — worth more than five minutes of description — and captures the address, access notes, and who will be home.

Then it acts. Connected to the shop's calendar, it offers real open slots and books the call. It sends the booking back in the same chat as a clean confirmation, and, where set up, takes a deposit or trip fee by payment in the chat — so the appointment is committed, not tentative.

The handoff is the point. The owner gets the full transcript plus a short summary: customer, problem, urgency, photos, address, what was booked, what was paid, anything that needs a human's judgment. And the agent hands off early by design: a true emergency, a firm price demand, an angry customer, or anything outside the written services list goes straight to a human, context attached. It books work. It does not negotiate, diagnose, or promise.

What you need to set it up

Where it breaks

Urgency can outrun chat. A flooded house needs a shutoff valve first, a conversation second. The agent must lead emergencies with the safety step and escalate fast — a calm, well-formatted chat with someone standing in water is a failure, not a feature. Gas smells, sparking equipment, anything life-safety: human and emergency services, never a booking flow.

Photos mislead. A stain looks like a leak, a leak looks like a stain, a model number in a blurry photo gets read wrong. The tech still verifies on arrival — the agent produces a starting point, not a diagnosis.

It quotes what it was told. If the website says free estimates and the owner no longer does them, the agent keeps offering them until someone updates the source. It learns the business as written, not as practiced.

Payments in chat settle arguments and start them. A deposit taken at 9 p.m. feels efficient until the job is cancelled or disputed — the refund path must be as clear as the payment was.

Not everyone is on iMessage, and not everyone wants a bot. Android customers and people who hate automated chat need a human one message away, not a maze.

Finally, transcripts nobody reads teach it nothing. The failures live in the conversations — the fumbled question, the job it should not have booked. Somebody has to review them, especially early on.

The bottom line

Run this if your shop's best marketing is a phone nobody can answer. The agent does not do the trade work. It does the first five minutes, instantly, at any hour, and hands you a booked, qualified, partly paid job with photos attached.

Start narrow: after-hours and overflow only. Read every transcript for two weeks. Widen it only when its summaries read like ones you would have written.

Related use cases

← All use cases

One working agent workflow, every day.

Real AI agent use cases from solar, construction, and field operations — what they do, what they cost, and where they break. No hype, no theory.