Booking confirmations and reminder messages

Customers who know what to expect cause fewer problems on the day. A short run of messages does most of that work. AI tools can draft them. You supply the facts.

The four messages

Confirmation. Sent when the booking is made. It repeats the date, time window, addresses, crew size, what's included and what the customer should do next. Keep it to the facts.

Access and prep. Sent a few days before. It asks about parking, lifts, stairs, narrow access, pets and anything not yet packed. This is the message that prevents surprises.

Reminder. Sent the day before. Short. Date, window, contact number, and the one or two things that matter most.

Arrival. Sent when the crew leaves for the first address, or when a delay happens. Keep it to where the crew is and when they expect to arrive.

What goes in each message

A good confirmation lists:

• Who is moving, and from where to where.
• The date and the arrival window, not a single time.
• What the booking covers, and what it doesn't.
• What the customer must do: pack, empty drawers, defrost the freezer, reserve parking, book lifts.
• How to reach you on the day.
• How to change or cancel the booking, using your actual terms.

Ask the tool to draft it from this list. Then check each item against the booking.

Text or email?

Text messages are short by nature. Wikipedia describes SMS as a service that lets mobile phones exchange short text messages.¹ That suits reminders and arrival notes. Longer content, such as prep lists, fits better in an email, with a text message pointing to it.

Put the date and window first, so they're seen even if the rest isn't read.

Make templates, not one-off drafts

Don't ask a tool to write each message fresh. Draft each message once and fix it by hand. Save it as a template with blanks for the details that change: [Date], [Window], [Pickup address], [Contact number]. Then the only job left each time is filling the blanks and reading the message back.

Some booking or messaging software can fill these blanks automatically. If you try one, ask:

• Can I see a preview of the exact message before it sends?
• What happens if a field is empty? Does it send a blank, or stop?
• Does it use the right time zone for each job?
• Can a customer reply, and who sees the reply?
• Is there an opt-out for customers who don't want texts?

Where AI drafts go wrong

Wikipedia says model responses can read well yet be "factually incorrect".² In a booking message the harm is quiet. Think of a swapped date or a time window nobody agreed. Or a pickup address mixed up with the delivery address, or an invented rule such as "no charge to cancel".

Read every message that contains a date, a time, an address or a fee as if it were a contract. For templates, check once carefully. For messages that vary, check each time.

Privacy and message rules

If you paste real booking details into a public AI tool, you're sharing personal information. Australia's regulator advises against entering personal information into public generative AI tools.³ A template with blanks avoids the problem, because the tool never sees the customer.

Rules on sending texts and emails to customers differ by country, so look at your own national regulator's pages on electronic messages and marketing. Keep reminder messages about the booking itself, not sales offers.

Start small

Pick the access and prep message first. It heads off the most surprises, so write it clearly. Then add the others. For how these messages connect to the rest of the office, see where AI fits.