Before You Build, Interrogate the Idea

One of the smartest things I have ever done in this whole process happened before I wrote a single prompt to build anything.

I opened up ChatGPT and asked it, essentially: is this idea any good?

Then I let it push back.

Then I let it push back some more.

Then I let it tell me the parts I didn't want to hear.

I lost some ideas that way. The ideas I lost were the ones that deserved to be lost. The ideas that survived that gauntlet were the ones actually worth building.

The free consulting firm in your pocket

Here's what most people miss: the same AI that will help you build the app is also, essentially, a research analyst, a market strategist, a business coach, and a devil's advocate — all willing to work for pennies at 2am. It's the closest thing to a McKinsey consultant we non-executives have ever had.

But almost nobody uses it for the pre-work. They jump straight to "help me build X" without ever asking "should I build X?"

Don't be that person.

What to interrogate before you build

Before you commit real time and money to a project, work through this list with the AI:

Problem

• Is this problem real for a specific group of people?
• How are they solving it right now?
• What do they currently pay to solve it?
• How annoyed are they about it, on a scale from "shrug" to "please help"?

Market

• How many people have this problem?
• Are they reachable? Where do they hang out?
• Is this a growing space or a shrinking one?
• Who else has tried to solve this? What happened?

Competition

• Who are the top three solutions right now?
• What do they do well? Where do they suck?
• What's my honest angle — cheaper, simpler, more focused, better for a specific niche?
• If a big player noticed my product, could they crush me in a week? Do I care?

Monetization

• How would this make money, specifically?
• What would a real person pay for this? Not "would they?" but "how much?"
• What's the cheapest version of this that could still charge for something?
• What are the actual costs to run it once it exists?

Feasibility

• Can this even be built with the tools I have?
• What's the hardest technical part, and does it require something I can't get through AI?
• What legal, regulatory, or trust issues am I ignoring? (This one is where a lot of ideas die.)

Honest gut check

• If I were reading about this in a newsletter, would I care?
• Am I excited about the problem or just the idea of being someone who built something?

Let it push back

The magic isn't in the answers. The magic is in the follow-up. When the AI gives you a comforting answer, push it: what would a critic say? When it gives you a critical answer, push it: what if I addressed that specific concern — does the idea survive?

The best pre-build conversations are essentially you and the AI trying to kill the idea. The ones that don't die are the ones worth building.

Real thing that happened to me

I had an idea I loved. I mean loved. I could see the app, the users, the story. Six paragraphs into an AI conversation about it, I realized the core assumption I was making about the audience was probably wrong. Two more rounds of questions and I was pretty sure the whole premise was shakier than I'd wanted to admit.

I would have found that out three months later, after spending real time and real money. Instead I found it out in an hour, for the cost of some electricity.

That single conversation was more valuable than most of what I built that quarter.

What not to expect from the AI here

• It won't tell you your idea is bad. It's too polite. You have to ask it to be honest.
• It won't have perfect market data. Cross-check anything that smells like a specific number.
• It won't know your local market, your specific network, your reputation. Bring those yourself.

But as a first-pass filter? As a way to catch obvious flaws before you fall in love? Unbeatable.

The takeaway

You have, sitting on your phone, the world's most patient sanity-checker. Use it before you build, not after.

The hour you spend interrogating an idea might be the most important hour of the entire project. It's certainly cheaper than the three months you'll save by not building the wrong thing.