How I Use AI Better Now

Over the weekend, I reviewed how I have been using ChatGPT.

The interesting part wasn’t how often I use it. It was how my usage has changed.
This week’s note is not about some magic prompts. I want to share the lessons I learned from this audit.

A few months ago, most conversations were simple:
I have a question → I ask →I get an answer.

I still use AI that way, but less often now.

More of my usage happens inside something I am already working on.

When I build an app, I use ChatGPT and Codex to think through the architecture, build features, debug problems, and review decisions.

When I write, I bring my own drafts, notes, published newsletters, and writing examples instead of asking AI to create everything from scratch.

I also give it more context: files, screenshots, previous discussions, and existing projects.

The biggest change is that these uses are starting to connect.

My old newsletter became material for an Idea Workshop. That Workshop gave me another newsletter idea.

My PKM thinking influences the apps I build. My writing helps improve the way AI edits my writing.

AI is becoming less of a place where I ask isolated questions and more of a layer within the work itself.

Here’s something you can try:

Review how you used AI over the past month.
Don’t just count how often you used it.
Look at what you actually did with it.

Which conversations helped you build something, make a decision, develop an idea, or improve your work?

Then look for repeated patterns.

You may find that the useful part isn’t a better prompt.
It is giving AI better context and keeping it involved in more of the workflow.

Maybe getting better at AI isn’t mainly about learning how to ask better questions.
It is about learning where AI belongs in the way you already work.

Until next time, keep learning.
Gav