One Prompt That Makes AI Brutally Honest

Most people use AI like a helpful assistant. I use it like a brutally honest advisor who owes me nothing.

Here’s the thing: we don’t need more validation. We need someone to tell us the truth. The problem is, friends are too kind, mentors are too busy, and advisors are too expensive. So I engineered a prompt that turns AI into the blunt, no-nonsense advisor I actually need.

I drop this in your Claude project instructions or cursor rules or at the start of every strategic conversation with AI:

By default, AI is agreeable. It validates your ideas, softens feedback, and wraps criticism in compliments. That’s useless when you’re trying to make hard decisions about your business, career, or life.

This prompt flips the dynamic. It gives the AI explicit permission to be direct, and in my experience, the quality of output changes dramatically. Instead of “that’s a great idea, here are some things to consider,” you get “here’s why this will fail, and here’s what you should do instead.”

I’ve used this prompt to pressure-test business strategies, review my marketing positioning, evaluate whether I’m solving the right problem, and gut-check decisions where I know I’m emotionally biased.

A few examples of what it caught that I was avoiding:

• I was building features nobody asked for instead of talking to customers

• My positioning was trying to please everyone and resonating with nobody

• I was spending time on technical optimisations when the real problem was distribution

None of these were comfortable to hear. All of them were true.

The value of AI isn’t in getting answers you like. It’s in getting the answers you need. Most people prompt AI to confirm what they already believe. The real leverage comes from prompting it to challenge what you believe.

Try it. Drop the above prompt into your next strategic conversation with Claude or ChatGPT. Ask it to review your business plan, your pricing, your career move, and your spending habits. Whatever you’ve been avoiding honest scrutiny on.

The worst that happens is you disagree with the feedback. The best that happens is you catch a blind spot that saves you months of wasted effort.

If you found this useful, I write about building in public, AI-assisted decision making, and the realities of building a Micro-SaaS as a solo founder. Subscribe for more.