Everyday AI: six scenarios you can use today

Reading about AI is interesting. Using AI today — on real chores — is what changes your week. Here are six scenarios I hand to every non-tech friend. Each takes under five minutes, and every prompt is ready to copy: just swap the bracketed parts for your own details.

Two words before we start

• Prompt — simply the message you send: your request plus the details around it. Not code, just text. Everything in the blocks below is a prompt.
• Follow-up — your next message in the same chat, the one that fixes the answer. Follow-ups are where a generic reply becomes yours, and they are the habit this whole article is about.

Anything in square brackets like this — [your text] — is a placeholder: delete the brackets and type your own thing there.

Scenario 1: the email you keep rewriting

You know how to say it out loud. But writing it down politely, briefly, and without sounding either rude or robotic — that’s the chore. Hand the tone work to the assistant:

Help me write an email.

Situation: [what happened and what I need]
Recipient: [boss / client / landlord / school teacher]
What I want to achieve: [apologize and reschedule / ask for a discount / say no politely]
Tone: [friendly but firm / formal / warm]

Write 2 versions: one shorter, one softer. Then ask me if anything sounds
not like me.

The last line matters. The assistant doesn’t know you — let it check whether the draft sounds like you, and correct it. That correction also teaches it your voice for next time — within this chat, or beyond it if your app’s memory feature is on.

Scenario 2: the trip or event plan

Planning is where assistants shine, because a plan is never final — you keep adjusting it message by message. Restaurant anniversary evening, a 4-day trip, a kid’s birthday for twelve kids indoors: same shape.

Plan [3 days in Krakow / a 40th birthday dinner / a weekend hiking trip].

My constraints: [budget], [who is coming], [things we love], [things we hate or can't do].
Ask me clarifying questions first. Then give me a plan with a plan B for
bad weather, and flag anything that needs booking in advance.

“Ask me clarifying questions first” is the small phrase that turns a generic itinerary into a plan for you.

Scenario 3: learning something new

The oracle mindset from the second lesson lives here. Instead of “explain machine learning” (too big), make it your personal tutor:

I want to understand [topic — e.g. how mortgages work / what a VPN does].

Explain it to me in three passes:
1. Like I'm 12 years old.
2. Like I'm a curious adult, with the real terminology.
3. The 5 mistakes people usually make about this.

After each pass, ask me one question to check I actually got it.

This is better than ten video tutorials, because it answers your questions instead of moving on without you.

Scenario 4: translation that keeps the meaning

Machine translation of single words has been solved for years. The value now is in tone and context — and the assistant can explain its choices:

Translate this to [language]: [your text]

Requirements: keep the tone (it should sound friendly, not stiff).
Then explain 2-3 places where you made a judgment call, and offer a more
formal alternative for each.

Scenario 5: meals for the week

The most unglamorous, most-loved scenario in my family. It helps you avoid buying food you won’t use, because you shop for what you’ll actually cook.

Make me a dinner plan for 5 weekdays.

We eat: [preferences and things we never eat].
Cooking time: max [30] minutes. I will shop once on Sunday.

Give me: the 5 dinners, one shopping list grouped by supermarket aisle,
and 2 planned leftovers so we cook once and eat twice.

Scenario 6: explaining a document

A lease, an insurance policy, a hospital report, terms and conditions. You are allowed to paste text like this into an assistant (just strip your passport number and other personal identifiers first):

Here is a document I don't understand: [paste the text]

Explain in plain language:
1. What is this document, in 2 sentences?
2. What am I agreeing to or promising?
3. What should worry me — the 3 riskiest points?
4. What questions should I ask before signing?

One rule for documents like these: a confident tone is not a check on facts. Ask the assistant to re-check anything doubtful, use it to prepare your questions — then confirm what actually matters with a professional or an official source.

The loop that makes any of these work

Every scenario above follows the same five steps. Click through them — this is the whole skill, and it works for every task you will ever give an assistant:

One conversation beats ten perfect prompts

Here’s the real habit. Watch how the follow-up does the actual work:

Press send next four times and watch the right-hand column: every message stays in the chat, so the third one can lean on the first. Notice what actually changed the result — not a cleverer first prompt, but one sentence of detail only you knew: a vegetarian in the family, 30 minutes, about 40 dollars. The first answer is a draft; the follow-up is where the plan becomes yours. Everything you and the assistant write in a chat is called its context — which is why the conversation, not the single magic prompt, is what carries the quality.

Do this today

Pick one scenario from above and run it in the next ten minutes. Not all six — one. The one whose chore annoys you most this week. The habit forms around relief, not around discipline.

1. Copy the prompt for that scenario into your assistant.
2. Fill in the brackets with your real details.
3. Send one follow-up that adds something only you know.

If you want to understand why some prompts work better than others, Prompt Anatomy takes the next step.