Morning-of and departure notifications that tell the customer who is coming, when, and what they look like, so nobody has to ring the office to find out.
An AI agent for technician arrival notifications is a 24/7 digital assistant that sends a morning-of schedule text, fires an on-the-way message the moment the previous job is closed out, includes the technician's name, photo and expected arrival, and answers the customer's follow-up questions in the same thread. It stops your dispatcher losing an hour a day to "where is he?" calls while the technician is already twelve minutes down the road. This is the cheapest reminder to run and the one customers notice most. A four-hour arrival window is an entire morning a homeowner cannot leave the house, so they ring at the halfway point to ask whether they can go and collect a parcel. Every one of those calls is four to six minutes of dispatcher time, and on a twelve-call day that is an hour of someone whose real job is booking work. The message costs nothing and removes the reason to call.
Tells the customer who is coming and when, before they think to ask.
Sends a morning-of text naming the arrival window and the technician assigned
Fires the on-the-way message the moment the previous job is marked complete
Includes the technician's first name and photo so a stranger at the door is expected
Gives a real ETA from the current route rather than a copied-and-pasted window
Answers follow-up questions in the same thread instead of routing them to the office
Flags a no-answer at the door so dispatch is told before the tech drives away
Count the inbound calls that exist only to ask where the technician is. Most shops running a dozen jobs a day get eight to fifteen of them, each running four to six minutes once the dispatcher has looked up the job, radioed the tech and rung back. That is close to an hour of a paid person, every day, producing nothing. The same message also protects the visit itself, because a customer who does not know when you are coming is the customer who nips out at exactly the wrong moment and turns a $110 job into a wasted roll.
A simple, three-step flow.
At a set hour the agent reads the day's board and sends every customer their window, the technician's name and a line about what the visit will cover.
When the technician marks the previous job complete or taps en route in the mobile app, the agent sends the on-the-way message with a current arrival estimate.
Questions about timing, parking and access are answered from an approved script. Anything about scope, pricing or a cancellation is routed to dispatch with the thread attached.
A realistic use case with concrete timing and output.
Scenario: An appliance repair company runs four technicians and about eighteen calls a day. The office logs roughly fourteen where-is-he calls daily, close to seventy minutes of dispatcher time. At 7:15am the agent texts each customer their window and technician. At 10:52am Devon closes out a dryer repair two suburbs over and taps en route, and the 11am to 1pm customer immediately gets: "Devon is on his way to you now and should arrive around 11:20. He drives a white van with our logo. Reply if you need him to use a side entrance." The customer replies asking him to park on the street because a delivery is due, and the note reaches Devon before he arrives. Across the day the office takes two timing calls instead of fourteen, and one customer who would have gone to the shops stays in.
Roles that gain practical value from this AI agent.
Gets back the hour a day currently spent relaying a technician's position between two phones.
Short jobs and tight routes mean arrival times move constantly, and customers notice when they do.
Long summer windows leave customers stuck at home, which is exactly when they start ringing.
Regular visits still need a heads-up so gates are unlocked and dogs are brought inside.
Fewer interruptions mid-job, and doors that open first time because the knock was expected.
A name and a photo in advance is a simple answer to real concerns about strangers at the door.
Key tools and what the AI agent does inside each.
Sends the morning schedule text and the on-the-way message, and carries the reply thread.
Provides the job status change that triggers the departure message and stores the customer's replies.
Signals when a technician marks a job complete or en route from the mobile app.
Supplies the day's booked windows for shops that schedule outside a full field-service platform.
Logs every notification sent and every reply so you can see which questions still reach the office.
Six practical scenarios that this AI agent excels in.
Common questions about using the AI agent in workflows.
An AI agent for technician arrival notifications is a 24/7 digital assistant that sends a morning-of schedule text, fires an on-the-way message the moment the previous job is closed out, includes the technician's name, photo and expected arrival, and answers the customer's follow-up questions in the same thread. It stops your dispatcher losing an hour a day to "where is he?" calls. Unlike a bulk reminder blast, it is triggered by real job status and it can hold a conversation about parking or access.
Answer the question before they ask it. Two messages do almost all the work: one in the morning with the window and the technician's name, and one at departure with a real estimate. Most shops running a dozen jobs a day take eight to fifteen of these calls, and the pair of texts removes the reason for nearly all of them.
The status change in your field-service software when the technician closes the previous job or taps en route. That is a far better trigger than a fixed clock time, because it reflects the day as it actually ran rather than the day as it was planned.
Only to the extent your field-service platform already does. If that platform exposes a route-based estimate, the agent uses it. If not, the message carries the technician's own estimate at departure. It does not add tracking you have not already set up.
Timing, parking, access and pets are answered from a fixed script you approve, and the answer is passed to the technician as a job note. Anything about diagnosis, pricing, warranty or canceling is escalated to dispatch with the whole thread, because the agent does not improvise on those.
Yes, and it is worth doing. A first name and a face turn an unexpected knock into an expected visit, which matters most for older customers and for anyone who has been warned about doorstep callers. You control which technicians have a photo attached.
The technician marks it on site and the agent texts and calls the customer immediately, then tells dispatch how long the technician is holding before moving on. That short window is often enough to catch someone who stepped out, which is exactly the trip you were about to waste.
Morning-of and departure notifications that tell the customer who is coming, when, and what they look like, so nobody has to ring the office to find out.