Start the day in your inbox and you have started it inside other people's priorities. By the time you surface it is 11 o'clock, you have been busy for 3 hours, and nothing you actually decided to do has moved.
A morning briefing is 5 minutes that happen before that. You hand over your calendar, your list and whatever is nagging at you, and you get back one screen: what matters today, what can wait, and the first thing to do when you stop reading.
The AI is not deciding your priorities. It is holding a mirror up to them, which on a foggy Tuesday is worth more than it sounds.
Most people who try this do it twice and stop, so the setup below is built for the version that survives. It takes 10 minutes once.
Set it up as a project, not a prompt
A project is a workspace with its own instructions and files. It costs nothing on the free plans of both major tools, and the reason to use one here is that a briefing is only as good as its context. In a project, the context is already loaded every morning instead of being re-pasted.
Make a project called Mornings. Put this in its instructions:
This project is for my daily morning briefing. I run [what your business does]. [Solo, or with staff, or with a partner.] My working hours are [when you really work, including if that is evenings around a job].
Every conversation here follows the same shape: I paste my day, you give me a briefing. Be direct. Do not cheerlead, do not pad, do not tell me I have got this. If something on my list has no business being there, say so. If I am overcommitted for the hours available, say that too rather than tidying it into a plan that cannot work.
Australian English. Keep briefings under 200 words unless I ask for more.
Then load 2 files if you have them: anything that lists your recurring commitments, and your current priorities for the quarter or the month. One page each is plenty. This is what lets it say "that is the third week this has moved" instead of treating every morning as the first.
The briefing prompt
Save this where you can paste it in one action. Your notes app, a text expander, wherever.
Here is my day. Calendar: [paste]. Task list: [paste]. On my mind: [anything nagging, including personal].
Give me a briefing: 1. The top 3 things that matter today, with one line each on why they earn the spot. 2. What on this list can wait or be dropped. Say it plainly. 3. Any collision or crunch in the calendar, including travel and anything that will run over. 4. The single first task to do the moment I stop reading, chosen so the day starts with momentum rather than dread.
Under 200 words. Do not soften the "can wait" list.
Paste the real calendar and the real list. Writing them out from memory defeats the point twice over: you will forget the thing you are avoiding, and the crunch warning needs actual times to spot an actual clash.
Include the personal. The load is one load. The invoice and the dentist appointment are competing for the same attention, and a briefing that pretends otherwise is planning a day you will not have.
Constructed example, a solo landscaper's Tuesday.
Top 3. 1) Ring the Nguyen job about the drainage variation, because the concrete is booked Thursday and it cannot be poured until they decide. 2) Send the Ballarat quote, it has been sitting 3 days and quotes go cold. 3) Order the turf for next week, the supplier needs 48 hours.
Can wait. The website photos, again. Also the tool shed tidy, which has been on this list for a month and is not a task, it is a mood.
Crunch. You have a 2pm on site 40 minutes away and a 3.30 pickup. That does not work with travel. Move the 2pm or move the pickup.
First task. Ring Nguyen. 2 minutes, and Thursday depends on it.
The version for a bad morning
The routine that only works on good days is not a routine. When you have 11 minutes and no capacity, use this instead:
No time this morning. Here is my calendar and list: [paste]. Tell me one thing to do today and one thing to consciously drop. 2 sentences. Nothing else.
It is not as good and it takes 90 seconds, which is the point. The habit survives the week that would have killed it.
What to do with the output, which is where most people fail
The briefing is not the routine. Reading a good briefing and then opening your inbox anyway is the most common way this dies.
3 rules that fix it:
Do the first task immediately, before anything else. Not after email. The whole design is that the day starts with a decision you made rather than a message someone sent you.
Actually delete what it told you to drop. If you leave the dropped items on the list, next week's briefing tells you to drop them again and you learn that the advice is ignorable.
Keep the briefings. Same project, so they accumulate. Once a fortnight ask:
Read back over the last 2 weeks of briefings in this project. What keeps appearing in the "can wait" list, what have I dropped and then re-added, and what pattern do you see in the days that went well against the days that did not?
That is where the compounding is. One briefing is a good morning. 20 briefings is a picture of how you operate, which nobody gets any other way.
Make this prompt yours
Paste your real day and this becomes this morning's briefing prompt. Nothing you type leaves this page.
How this routine fails
It cannot see your calendar unless you paste it or connect it. If you have connected your calendar, check what you have granted. If you have not, pasting takes 10 seconds and is the safer default.
It does not know what is urgent in your industry. It is reading your list, not your context. If something looks droppable to it and is not, say so and it will remember within that project.
It will make you feel productive without you doing anything. 5 minutes of planning theatre is worse than no planning, because it feels like progress. If your briefing routinely takes more than 5 minutes or runs past 200 words, it has become the work instead of the thing that starts the work.
It is not a therapist and the "on my mind" field is not a journal. 2 lines. The point is to get the nagging thing out of your head and into the sort, not to process it.
Run it tomorrow morning
Set up the project and run it once, tomorrow morning, with your real calendar. One run tells you more than any amount of reading about whether this is for you.