Stop querying. Start briefing.

You're probably using ChatGPT wrong

Author

Kyle Balmer
April 07, 2026

I train companies and individuals on how best to use AI.

Here's a thing I've seen a hundred times. Maybe more.

Someone tells me they "use AI". I ask how. They open the app, show me their recent chats, and it's fifty-odd conversations that all look the same. A one-line question, a decent-ish answer, move on. Next one-line question in a brand new chat. Rinse, repeat.

That's not really using AI, that's using a slightly friendlier search box.

Level 1 of the AI skill tree is really a habit shift.

We need to stop using AI as a Google replacement.

We don’t just use AI to answer questions. Instead we use AI to ask better questions.

The difference is giving AI enough context to actually help you, rather than firing off questions and hoping. One proper brief beats twenty one-liners every time. Most people never learn this and then wonder why their AI output feels average.

Honestly this is why a lot of people bounce off AI. They use it in this sorta naieve way. Get fairly unexciting results. And stop.

Entirely fair. But it’s also (without being too mean) user error.

I’ve broken Level 1 up into a handful of subskills:

Level 2 is about taking our best practices and making sure we don’t repeat ourselves.

Ever gone into ChatGPT to do daily task and opened up a new chat? Don’t do that! Sure it works but it’s making your life harder than it needs to be - embrace laziness! Sorry, efficiency.

Every major AI tool now has a persistent workspace feature: ChatGPT Projects, Claude Projects, Gemini Gems (for some reason Google still don’t have proper Projects…comeon, guys).

You can give one recurring job a permanent home, feed it the documents and context that matter once, and come back to it all week. No more starting from scratch each time.

If you skipped the placement test yesterday, Level 1 is where most people should start regardless. It's short, practical and lays in some good habits that make everything else makes sense.

Tomorrow: the two levels where AI stops feeling generic. We’ll make it your AI.

Kyle