Prompt: the four-part prompt for Claude (Context, Goal, Example, Rules)
A fill-in-the-blanks prompt for any everyday task. Tested October 19, 2026 on Claude Sonnet 5.5 against lazy prompts on 5 tasks: 15/15 vs 11/15 on right facts, nothing made up and ready to use.
Works with: Claude, ChatGPT
The task
Get an answer you can use as-is for an everyday task: slides, a to-do list, a plain-words explanation, a polite reply.
Context: [WHO YOU ARE AND WHAT'S GOING ON, IN ONE OR TWO LINES] Goal: [WHAT YOU WANT AT THE END] Example: [ONE LINE THAT LOOKS LIKE A GOOD ANSWER] Rules: [WHAT IT MUST KEEP, WHAT IT MUST NOT ADD, HOW LONG] [PASTE YOUR NOTES, EMAIL OR LETTER HERE]
How to use it
- Context: say who you are and what is going on, in one or two lines.
- Goal: say what you want at the end.
- Example: write one line that looks like a good answer.
- Rules: what it must keep, what it must not add, how long. Then paste your text.
What you get
An answer that keeps every date and fact, adds nothing you didn't give, and needs no edits. In our test the rules did the most work: 'add no places, prices, bookings or activities' turned a 300-word invented slide deck into 3 clean slides.
Here's the text for your three slides: **Slide 1 - Friday** - Train leaves at 9:10 a.m. **Slide 2 - Saturday** - Museum at 2 p.m. **Slide 3 - Sunday** - Train home at 5:30 p.m. I used one bullet per slide because you gave me one fact per day, and I didn't want to pad the slides with anything you didn't say. If you want more on screen, send me more details, like what you'll do Saturday morning, and I'll add them.
Go further
- How to use Claude AI for free: we tested cheat-sheet prompts on 5 everyday tasksWhat Claude's free plan includes, where to click, and the four-part prompt (Context, Goal, Example, Rules) that scored 15/15 in our October 19, 2026 test on Claude Sonnet 5.5. Lazy prompts scored 11/15, and 'world-class expert' filler made things worse.Read the article
- Prompt: fact-check a ChatGPT answer against a source you trustPaste an AI answer and a source you trust. The AI splits the answer into claims and labels each VERIFIED, WRONG or CAN'T VERIFY, with an exact quote. Tested October 18, 2026 on Claude and GPT-6.1: both caught a planted wrong date.See the prompt
Where should you start with AI?
The free assessment gives you, in 4 minutes, the 3 tasks to automate first in your business.