Prompt Like A Pro

Most people use ChatGPT wrong.

Everyone seems like an “AI expert” these days.

So I asked Claude’s latest model to analyze 1,500 academic papers to find what actually works.

No fluff. No theory. Just data-backed results.

Here are the 12 prompting laws that have elevated my game 👇

1/ Include the context
→ First 2 lines = background + goal (AI needs both)

2/ Force structured thinking
→ “Let’s solve this step-by-step” unlocks better logic

3/ Ask for specifics
→ “500 words” beats “short post” every time

4/ Generate then curate
→ Ask for 3 versions, steal the best bits from each

5/ Give 2-3 examples (not more)
→ 2 examples > 10 examples (research-backed)

6/ Switch perspectives
→ Customer POV + Expert POV = fuller picture

7/ Push for better
→ “How would you improve this?” = instant upgrades

8/ Role-play strategically
→ Characters for stories, direct asks for facts

9/ Frame it positively
→ “Make it punchy” > “Don’t be boring”

10/ Build on wins
→ Good answers spark better questions

11/ One task at a time
→ Single focus = sharper output

12/ Repeat what matters
→ Key details twice = better results

Mastering these prompting laws?
That’s step 1.

But the reality is that, 90% of AI content fails because of 𝘸𝘩𝘢𝘵 you write, not 𝘩𝘰𝘸 you write it.
The topics, the tone, and more.