Stop Asking AI Questions. Make AI Ask You Questions Instead.

A simple tactic that dramatically improves AI output.

In Between Meetings & Meals 004, I talked about something I call building the bones first.

The idea is simple. Before you ask AI to create something meaningful, you give it structure. The framework. The context. The bones of what you’re trying to build.

Once those bones exist, AI has something real to work with.

But there is another move that takes the output to the next level.

If building the bones gives you structure, this tactic gives you detail.

It fills in the nuance.

And the best part is how simple it is.

All you have to do is ask.

Most people open AI and start asking questions.

“Write this.”

“Help me with that.”

“Fix this email.”

That works, but it often leads to answers that feel generic.

The upgrade is this:

Ask AI to ask you questions.

Instead of guessing what context the system needs, you let it pull the information out of you.

For example, instead of saying:

“Help me build a framework for this project.”

You can say:

“Before you answer, ask me 20 questions that would help you understand this better.”

When you do that, something interesting happens.

AI starts asking about things you may not have thought of yet.

The audience.
The timeline.
Constraints.
Stakeholders.
Success metrics.

Now you are not just throwing a prompt into the void. You are building the input together.

The more context you feed it, the better the output becomes.

This is one of those small tactical moves that dramatically improves the results.

It is also where I discovered a little shortcut about two years ago.

The slowest part of answering all those questions is typing.

You can think faster than you can type, and the process starts to feel jerky when you are trying to keep up.

So when I get to this part of the process, I do something different.

If I am working at my computer, I pull out my phone and turn on voice to text. You can also just turn on the microphone on your laptop now.

Then I start answering the questions one by one out loud, almost like I am being interviewed.

Instead of slowly typing everything out, I just talk.

The answers come faster. The thoughts are more natural. And the context builds quickly.

By the time AI processes those responses, it now has layers of nuance that were never in the original prompt.

That is usually when the output levels up.

This move works whether you are:

building a framework
solving a problem
writing something important
learning a new skill
or designing a system

First you build the bones.

Then you let AI ask the questions that fill in the details.

It is a simple shift, but it completely changes how the system works with you.

It turns AI from something that just answers prompts into something that helps you think.

This is one of the tactical ideas I talk about in The AI Toolkit.

The AI Toolkit

I wrote the book because most people are still using AI for surface-level tasks when it can actually help you build frameworks, solve complex problems, and learn new skills much faster.

If you are curious about the prompts and workflows I use every week, you can check it out there.

But for now, try the move.

Next time you open AI, start with this:

“Before you answer, ask me the questions you need to understand this better.”

Then just start answering.

You might be surprised how much better the results get.