Skip to content

Reminder Messages for Deadlines and Monthly Cut-Offs

Last checked 4 November 2026

Never ask an AI tool for a deadline. Give it the date and ask it to write around it.

That's the whole method. The rest of this page explains why it works and how to put it into practice with a regular monthly message that clients actually open.

Why dates are the risky part

A language tool writes by producing believable text. It has no calendar of your clients' obligations, and it can't tell you what changed this year. When it states a date from memory, the date can be wrong and still read well.

Wikipedia has a name for this kind of mistake: a hallucination.1

A wrong date in a reminder does more harm than no reminder, because the client trusts it and plans around it. Check twice.

Two kinds of date

Your own cut-offs. The day you need bank statements for the month. The day you stop taking new documents before a report. The day you'll call if nothing has arrived. You set these, so you know them.

Official deadlines. Filing and payment dates set by a tax authority or other body. Don't write these from memory or from a tool. Look at the official page, copy the date, and name the source in the message if you can.

The Australian Taxation Office warns that some of its information applies to one financial year. It says to confirm you have the right year's information first.2 The same habit helps anywhere. Check the year as well as the day.

If a client asks what a deadline means for them, send them to the authority or to their registered adviser. A reminder is a nudge, not advice.

How a reminder gets its datesTake dates from your own calendarCheck official dates at the sourcePaste them into the promptRead the draft against the datesSchedule or send it yourself
Dates go in from your records and come out checked.

What makes a reminder get read

The US government's plain-language guidance recommends a clear topic sentence and active voice.3 For a reminder, that means a few habits. They're quick.

  • Put the action and date in the subject line
  • Open with what you need, not with pleasantries
  • Name the one thing they need to do
  • Give a single way to reply
  • Keep it to a few lines

Compare "A gentle reminder regarding upcoming submission requirements" with "Please send May receipts by Friday the fifth". The second tells the reader what to do before they've finished the first line.

Building a monthly reminder

Give the tool a pattern, with the dates already filled in. Wikipedia notes that a prompt can carry sample text for the model to copy from.4 So paste in one reminder you like and ask for the same style.

Here is a monthly reminder I like: [paste your own]. Write this month's version in the same tone. The cut-off for documents is [date, day and year]. Mention that anything after that goes into next month. Keep it under eighty words. Don't change any date.

Then check it against your calendar. Look at the day of the week as well as the number. Tools sometimes pair a date with the wrong weekday.

A rhythm that works for most practices

  • One message early in the cycle with the cut-off date
  • One short nudge a few days before it
  • One message on the day for anyone who hasn't sent anything

You choose the gaps. Draft all three in a sitting, check the dates once, and schedule them in your email programme. The draft of the second and third can be shorter than the first.

Keep clients' details out of it

A reminder usually needs no personal information at all. Use a template with "[Name]" and merge in names inside your own email tools. Before you paste any real client information into a chat tool, read client data and AI tools.

When a reminder turns into a chase

If the cut-off passes with nothing in, switch to a missing-documents message. That one names the items and offers a way forward, which a general reminder doesn't.

Sources used

  1. Hallucination (artificial intelligence), Wikipedia
  2. Record keeping for business, Australian Taxation Office
  3. Principles of plain language, Digital.gov (U.S. General Services Administration)
  4. Prompt engineering, Wikipedia