30 ChatGPT Tips That Actually Change How You Work

Most people use it like a search engine. That’s why it disappoints them.

Mar 30, 2026

∙ Paid

Over 200 million people open ChatGPT every week and wonder why it feels mildly useful.

That’s not a ChatGPT problem. That’s an interaction problem.

Horizontal comparison infographic showing how 200 million weekly ChatGPT users use it wrong versus right. Left dark panel shows the wrong approach: no context, no role assigned, accepting the first draft, editing heavily, starting from zero every session, never specifying what to avoid, no writing examples, no format specified. Right panel shows the correct approach: assign a specific role first, give full context upfront, make it ask you questions, paste your own writing as a style reference, specify format before it writes, list what to never say, run 5-round iteration loops, and paste a 150-word context block every session. Three insight cards at the bottom summarize: without these habits output is generic and heavily edited, with these habits output is specific and on-voice, and the real difference is habits not intelligence. — 200 million people open ChatGPT every week. Most treat it like a search engine. A few small changes to how you prompt and give context, and the tool feels completely different. Source: The AI Corner, 2026.

A few small changes to how you prompt and how you give context, and the tool feels completely different. These are the tips I actually use:

1. Give it a role before you give it a task

“You are a senior B2B copywriter who has worked on SaaS product launches. Your writing is direct, specific, and avoids marketing jargon.”

That one sentence changes the voice, the structure, and the vocabulary of everything ChatGPT produces after it. The more specific the role, the more specific the output.

2. Describe what you want to avoid

Add a short list of what it should never do:

“Never use the words leverage, streamline, or cutting-edge. Never open with a question. Never write a list when a paragraph will do.”

This one addition eliminates most of the outputs you’d otherwise spend time editing manually.

3. Paste in examples of your own writing

ChatGPT matches a style it can see. It cannot match a style you describe.

Paste two paragraphs of something you’ve actually written. Then say: “Match the tone, sentence length, and vocabulary of this. Use it as a reference, not something to copy.”

This works better than any description of your voice you could write.

4. Make it ask you questions before it writes anything

“Before you write anything, ask me the five to ten questions that would most improve what you produce. Do not start writing until I have answered them.”

The questions it asks will surface things you forgot to specify. The output after this step is almost always stronger than anything you’d get from a direct prompt.

5. Build a context block you paste every time

Write this once. Paste it at the start of every important conversation.

I am [role]. I work on [what you do]. 
My audience is [who they are].
My writing style: [3-4 specific words].
Words I never use: [your list].
Format I prefer: [your default].

Every session starts with full context instead of zero. Takes 90 seconds to write once. Saves time forever.

The other 25 are for premium subscribers

The ones below are the prompting patterns that produce the biggest jumps in output quality. Chain-of-thought triggers, steelman prompts, pre-mortem templates, iteration loops, assumption audits, and 20 more.

As a premium sub you also get every issue ever published in The AI Corner. 60+ guides, all of it immediately.

Start your free 7-day trial

Cancel anytime. First subscribers get 50% off forever.

the premium 25 ChatGPT Tips:

6. The chain-of-thought trigger

Add this to any prompt where you want ChatGPT to reason carefully rather than pattern-match:

Keep reading with a 7-day free trial

Subscribe to The AI Corner to keep reading this post and get 7 days of free access to the full post archives.