Brain-dump like you're texting a friend

The situation

It’s 9.40pm. Forty things are rattling around your head: the permission slip, the dentist, the birthday present, the thing you promised your sister. You open an AI app, see the empty box, and freeze, because surely you’re meant to type a clever, perfect “prompt” (the message you give the AI). You’re not. Messy works better.

By the end of this lesson you’ll know, in plain English, what an AI chat app is good at and where it trips up. You’ll have one installed, and you’ll have turned a head full of noise into a sorted list in about two minutes.

Why it’s worth it

The “mental load” is the invisible job of keeping track of everything: noticing the shoes are too small, remembering the form is due, knowing the party is on Saturday and that a present needs buying before then. None of the tasks is hard on its own. What wears you down is carrying all of them at once, all day, and re-sorting them in your head every time something new lands.

Writing things down helps, but sorting the list is the tiring part: what’s urgent, what can wait, what someone else could do. An AI app is surprisingly good at exactly that first pass. It doesn’t judge the mess, it doesn’t get tired at 9.40pm, and it takes seconds. You stay the one who decides. It just hands you the list in a shape you can act on.

How an AI chat app actually works

You don’t need any technical knowledge for this course, but one minute on how these apps work will make every later lesson make sense.

An AI chat app (ChatGPT, Claude or Gemini) has learned from an enormous amount of writing: books, articles, websites, conversations. From all that, it has become very good at language. It can tell what you mean from a rambling, misspelled message, and it can rewrite things in a different tone, sort a jumble into groups, summarise a long letter or draft an email. When it answers, it’s writing what fits best given everything you’ve said so far.

That’s also where its weak spot comes from. It doesn’t know your life. Unless you tell it, it doesn’t know your daughter’s dentist, the school’s term dates or that Saturday’s party starts at two. When a piece of information is missing, it doesn’t stop and say “I don’t know” as often as it should. It fills the gap with something that sounds likely, and says it confidently. Tech people call this “hallucinating”. In this course we’ll call it confident guessing, and it’s the reason every lesson ends with a quick check.

Two more things worth knowing. Within one chat, the app remembers what you’ve said, so you can keep refining without repeating yourself. A new chat starts from a blank page, apart from any settings you’ve saved (Lesson 1.3) or a memory feature you’ve switched on. And some apps can search the web, but even then they can mix things up, so facts that matter still get checked against the real source.

An AI app is great at	It’s weak at
Understanding messy, everyday language	Facts you haven’t given it
Sorting, summarising, rewording	Dates, times and numbers it has to work out
Changing tone: angry to polite, long to short	Knowing your family, your school, your rules
Drafting messages and lists in seconds	Saying “I don’t know” when it should

Pick your app

This course works with any of the three big AI chat apps: ChatGPT, Claude and Gemini. All three have a free plan and an app for iPhone and Android, and every prompt in these lessons works in all of them. They differ in personality and in where the buttons are, not in whether they can do this course.

If you already have one, use that. Seriously, that’s the whole decision. If you’re starting fresh, pick one, download it from your phone’s app store and sign up with your personal email or account. Give it a week before you judge it: the first few answers often feel a bit generic until you’ve learned how to talk to it, which is what this module is for. The free plans are enough to follow this course; paying mainly buys higher limits and the newest models.

Before you put anything personal into it, Lesson 1.4 covers what to keep out and which privacy settings to switch on. For today’s brain-dump, everyday to-dos are fine.

Why messy works better than tidy

Most people’s first instinct is to write something short and sensible, like “Help me organise my week.” It feels like the right way to talk to a computer. But think about what the app gets from that: nothing about your week. So it guesses, and you get a generic time-management article with tips about colour-coded calendars.

Now compare a messy dump: “school photo money?? thurs i think. mia dentist needs booking. present for leo’s party sat.” Every fragment carries information: a deadline, a name, a doubt (“i think”). The app reads meaning, not grammar, so typos and missing punctuation cost nothing. The more of your real situation it has, the less it has to guess, and the less it guesses, the more useful (and correct) its answer is.

So the rule for this whole course is: give it everything, in whatever order it comes out. Tidying up is the AI’s job.

Copy this

Copy the prompt below into your AI app. Replace the bit in [square brackets] with everything on your mind: appointments, forms, birthdays, worries, half-finished jobs. No order, no punctuation, no tidying. Then send it.

I'm going to brain-dump everything on my mind. It's messy, that's fine.

[Everything on your mind goes here]

Now sort it into four groups:
1. Must happen today
2. This week
3. Can wait
4. Someone else could do this (say who, if I mentioned them)

Keep my words where you can. Don't add tasks I didn't mention. If something is unclear or might have a deadline, ask me up to 3 questions at the end.

Why each part is there

“It’s messy, that’s fine” tells the app not to comment on the mess or ask you to rephrase. It just gets on with it.

The four groups give it a shape to sort into. Without them, it would invent its own categories, often clever-sounding ones (“Health”, “Social”, “Admin”) that don’t help you decide what to do tonight. “Someone else could do this” is there on purpose: a surprising amount of the mental load is jobs that don’t have to be yours.

“Keep my words” stops it rewriting “call Mum back” into “Re-establish contact with your mother”. Your own words are faster to recognise when you glance at the list.

“Don’t add tasks I didn’t mention” is the guard against confident guessing. Without it, a list that mentions a school trip can sprout a “pay deposit” you never said.

“Ask me up to 3 questions” gives it a way to say “I’m not sure” instead of guessing. The limit of three keeps it from interviewing you.

What you’ll see: after a few seconds, the app replies with your four groups and a couple of questions underneath.

What you’ll get

Here’s a real-life dump, typed one-handed:

school photo money?? thurs i think. mia dentist needs booking. sam’s football boots too small. present for leo’s party sat. call mum back. excursion permission slip. bins. dishwasher tablets. work report friday. haven’t replied to class chat about bake sale

And the kind of answer that comes back (an example: your app will word it differently, and that’s fine):

Must happen today

• Sign the excursion permission slip
• Call Mum back

This week

• School photo money (you think it’s due Thursday)
• Book Mia’s dentist appointment
• Present for Leo’s party on Saturday
• Work report, due Friday
• New football boots for Sam

Can wait

• Dishwasher tablets

Someone else could do this

• Bins
• Reply to the class chat about the bake sale (a quick “can’t help this time” is fine)

Questions: When is the permission slip due? Is the photo money definitely due Thursday? Do the boots need to be bought this week, or can it wait until the weekend?

Notice three things. It kept your doubt: “you think it’s due Thursday” instead of turning a guess into a fact. It turned fragments into clear tasks you can tick off. And its questions are exactly about the things it couldn’t know: deadlines. That’s the prompt doing its job.

Notice too what it got slightly wrong: it put the permission slip under “today” without knowing the deadline. That’s a reasonable guess, but it’s still a guess, which is why you’re the one who has the final say.

Before you rely on it

You’re not sending this list to anyone, but give it ten seconds before you trust it:

• Did everything make it? Glance back at your dump. In long lists, an item occasionally gets dropped or merged into another one, because the app is summarising, not copying.
• Did anything appear that you didn’t say? A deadline, a task, a detail. That’s confident guessing. Delete it.
• Is “today” really today? The app is reading urgency from your words, and it doesn’t know your week. Move things around: you know what’s really urgent, it doesn’t.

Make it yours

Keep going in the same chat, because it already holds your whole list. Each of these takes one line:

• Which three things would make tomorrow easier if I did them tonight? Useful when the list is long and you’ve got twenty minutes of energy left.
• Turn "this week" into a checklist I can paste into my notes app. Plain checklists paste cleanly into most notes and reminders apps.
• Write a short message to [partner's name] asking them to take [tasks]. Friendly, not naggy. The “someone else” group is only useful if someone else hears about it.
• Fit the "this week" items around [my work hours, the school run, Saturday sport]. Turns a list into a rough plan.

If it goes sideways

• It wrote an essay. These apps tend to over-explain unless told otherwise. Reply “Shorter. Bullet points only.” Lesson 1.3 makes that the default so you never have to ask again.
• It added self-care advice you didn’t ask for. Many AI apps are cautious and caring by default. Reply “Skip the advice, just sort the list.”
• It got something wrong. Correct it in plain words: “No, the party is Leo’s, not Mia’s.” There’s no special command; it understands corrections the same way it understood the dump.
• It says you’ve reached a limit. That’s the free plan’s limit, not a fault. Try again later, or paste your dump into another app.
• After a few weeks, the list gets muddled. Very long chats are hard work for AI apps: details from the start can get lost or mixed up. Ask it to “list everything that’s still open”, then paste that list into a new chat and carry on from there.