Blog

The AI interview method: sharpen any idea with 5 questions

Here's a small habit that will change how you think: stop asking AI to summarize your idea, and start letting it interview you. A summary hands your own half-formed thought back to you in tidier words. An interview does something far more valuable — it draws out what you haven't figured out yet. That's the difference between a note-taker and a thinking partner.

Why interviewing beats summarizing

When you explain an idea out loud, you fill gaps with confidence — "it's like X but better," "people will obviously want this." A summary preserves those gaps because it only reflects what you already said. A good question does the opposite: it pokes exactly where you were vague. "Better how, specifically?" "Which people, and what were they doing before?" Answering those is uncomfortable, which is the point — the discomfort is you discovering the parts of the idea you hadn't actually thought through.

You don't understand an idea until something has asked you the questions you were avoiding.

The 5 questions

You can run this method by hand — with a notebook, a friend, or an AI you prompt to "interview me, one question at a time, about this idea." Here are the five that do the most work, in order:

1. Who is this really for?

Not "everyone." Name a specific person in a specific situation. If you can't, the idea isn't wrong — it's just not ready, and this is the gap to close first.

2. What are they doing today instead?

Every idea competes with a current workaround, even if that workaround is "nothing." Naming it tells you what you're really up against and how much better you'd have to be to matter.

3. What's the honest why?

Strip the pitch. What painful thing does this remove, or what valuable thing does it add, for that specific person? If the honest answer is "it'd be kind of neat," you've learned something important early and cheaply.

4. What has to be true for this to work?

Surface the riskiest assumption hiding underneath — the thing that, if false, sinks the whole idea. Usually it's about demand, cost, or distribution. Now you know exactly what to test first.

5. What's the smallest first step?

End every interview here. What could you do in two minutes to move this forward? An idea that ends with a concrete next step is an idea that survives; one that ends with "interesting!" is one you'll forget by Thursday.

Try it right now

Take any idea rattling around your head and answer those five out loud. Most people find that by question three the idea has either gotten noticeably sharper or quietly revealed a fatal flaw — and both outcomes just saved you time.

Built into an app

Doing this by hand works, but it takes discipline you won't always have at the moment an idea strikes. That's the exact job of Clarify in Thought Spark: it runs this interview for you on any captured idea — asking the right questions, tuned to how you think, and handing back a structured, sharp version you can then research and turn into action. It's the interview method, minus the willpower. For the bigger picture, read how to turn ideas into action.

Keep reading

← All posts