Claude General

Recommendation Prompts: 6 for Checking Whether the Answer Is Actually Yours

ChatGPT is adding Sponsored Agents, where clicking an ad opens a labelled conversation with a business. Labelling solves the obvious problem and not the harder one — whether the unsponsored answer changes because sponsorship exists. These six prompts are for getting a recommendation you can trust, from any assistant, regardless of how it is funded.

⌨️ 6 prompts 🕐 Updated Sep 18, 2026

💡 How to use these prompts: Replace everything in [BRACKETS] with your specific details before sending. Click Copy to copy any prompt to your clipboard instantly.

Ask for the case against your shortlist

Asking who a product is wrong for surfaces more than asking what is wrong with it.

I am choosing between: [LIST THE OPTIONS]

For each one, tell me who it is wrong for and why. Not the weaknesses in general - the specific person who would regret buying it.

Then tell me which of those descriptions sounds most like me based on what I have told you.

Make it name what it does not know

Recommendations age badly. Pricing and limits are the first things to go stale.

You recommended: [PASTE]

Tell me:
- What you know about this from documentation versus general reputation
- When your information about it is from
- What could have changed since
- What you would check before relying on this

If pricing or features are involved, say plainly that they change often.

Force a recommendation you did not expect

The obvious answer is obvious because it is marketed well. This surfaces the alternative.

Give me your recommendation for: [DESCRIBE THE NEED]

Then give me the option most people in my position overlook, and explain why it gets overlooked.

If the overlooked one is actually better for me, say so.

Separate the popular from the suitable

The most-recommended tool is often the best-funded one, which is a different thing.

For this need: [DESCRIBE], tell me:
- What most people choose
- What an expert in this area would choose
- Whether those differ, and why
- Which one fits my situation

Do not assume popular means correct.

Cross-check one assistant against another

Two assistants with different funding models disagreeing is more informative than either alone.

Here is a recommendation I got from a different AI assistant: [PASTE]

Assess it. Where do you agree, where would you differ, and what did it leave out?

Be specific about disagreements rather than diplomatically endorsing it.

Ask what would change the answer

Knowing what would change the answer tells you how much the answer depends on what you said.

You recommended [X] for [MY SITUATION].

What would have to be true about my situation for a different option to win? Give me the three facts that would flip it.

If none would, say the recommendation is robust and why.