10 prompt rewrites, before and after

These are the kinds of prompts people actually type, and what they look like once they stop making the AI guess. None of the "after" versions are clever. Most just add the reason and one limit.

Copy any of them and swap in your details.

1. The email to your landlord

Before: write an email to my landlord about the heat

After: Write a short, polite email to my landlord. The heat in my
apartment has been out since Tuesday and I reported it by phone on
Wednesday. Why: I want it fixed this week and I want a written record
that I asked. Keep it calm, under 120 words, and ask for a date.

The "why" is doing two jobs here. It tells the AI the tone (calm, not angry) and it tells it to include the dates, because a record needs dates.

2. Understanding a bill

Before: explain this bill

After: Here's my phone bill [paste]. Explain each charge in one plain
line. Why: it went up $14 this month and I want to know which line
changed and whether I can turn it off.

3. A week of dinners

Before: make me a meal plan

After: Plan five weeknight dinners for a family of four. Two of us
don't eat pork, one kid won't eat anything spicy, and I have about 30
minutes after work. Why: I shop once on Sunday, so give me one combined
grocery list at the end, grouped by aisle.

The grocery list is the part people forget to ask for, and it's the part that actually saves the time.

4. Feedback on your own writing

Before: is this good?

After: Read my cover letter below as a hiring manager with 80
applications to get through would. Tell me the one paragraph you'd skip
and why. Don't rewrite it. I want to fix it myself.

"Is this good?" almost always gets "Yes, with a few suggestions!" Asking what a tired reader would skip gets you the truth. And "don't rewrite it" keeps the AI from handing back something that isn't yours anymore.

5. A toast

Before: write a wedding toast for my brother

After: Help me write a 2-minute toast for my brother's wedding. I'll
give you three true stories below. Why: I'm nervous speaking and I want
something I can say without reading it word for word, so keep the
sentences short and give me a one-line outline I can memorize.

You bring the stories. The AI is great at shape and terrible at making up memories, and a toast with invented memories is worse than no toast.

6. A tricky reply to a customer

Before: reply to this angry review

After: Draft a reply to the review below. Why: the customer has a point
about the wait, but the refund they mention isn't something we offer.
I want to own the wait, explain the policy once without arguing, and
invite them back. Under 80 words. No "we're sorry for any
inconvenience."

Banning that one phrase changes the whole reply. It's the line everyone recognizes as a template.

7. Studying for a test

Before: explain photosynthesis

After: Quiz me on photosynthesis for a 10th-grade biology test. Ask me
one question at a time and wait for my answer. If I'm wrong, explain it
in two sentences and ask a follow-up. Why: I understand it when I read
it but freeze on the test, so I need practice answering, not more
reading.

8. A bug in your code

Before: why doesn't this work [paste code]

After: This function should return the total of a cart, but it returns
NaN when the cart has a discount item. Here's the function and one
example input that fails [paste]. Find the cause before suggesting a
fix, and change as little as possible.

"Change as little as possible" is the line I'd put in nearly every coding prompt. Without it you can ask for a one-line fix and get back a rewritten file.

9. Summarizing something long

Before: summarize this

After: Summarize the document below for someone deciding whether to
read it. Why: I'm sharing it with my team and most of them will only
read your summary. Five bullets, then the one question it leaves
unanswered.

Asking for what's missing is the trick here. Summaries tend to make documents sound more complete than they are.

10. Planning a trip

Before: plan a trip to Chicago

After: Plan two days in Chicago for two adults who like food and
walking, and don't care about shopping. We arrive Friday at noon and
leave Sunday at 4. Why: we've been once before and did the big
tourist stops, so skip those. Keep each day to three main things,
with walking times between them.

"We've been before, skip the big stuff" cuts out half of the generic answer before it's written.

What these have in common

Look back at the afters and the same few moves keep showing up: the real situation, the reason, one hard limit, and sometimes a phrase to avoid. That's really it. I wrote more about the reasoning in how to write a prompt that works the first time, and if you're writing anything personal, how to get AI writing that sounds like you picks up where #4 and #5 leave off.

If you'd rather not type them out, NeLocker has free prompts and fill-in-the-blank templates for most of these already. The beginner's guide to prompting is there if you want the basics first.

Quick answers

What makes a good AI prompt different from a bad one? A good prompt says what you want, who it's for, why you want it, and any hard limit like length. A bad one leaves the AI to guess all four.

Do I need to give the AI a role, like "you are an expert"? Rarely. Telling it the situation and the reason does more than a job title. Use a role only when the perspective itself matters, like "read this as a nervous first-time buyer would."

Can I reuse these prompts? Yes. Copy the "after" version, swap in your own details, and save the ones you use often so you don't rewrite them every time.