Getting useful AI help
Set comparison criteria before asking for a recommendation
Write the conditions an option must meet, the preferences that matter and anything that rules it out. Ask for a comparison against that list, with unknowns shown clearly. Keeping the criteria visible lets you see why a recommendation follows and whether the same standard was applied to every option.
A text to start with
I am choosing between [options or type of option] for [purpose]. My must-haves are [conditions]. My preferences are [preferences], in this order of importance: [order]. These issues rule an option out: [disqualifiers]. Compare the options against only these criteria, show where each one fails, and recommend one. Do not add criteria of your own.
Example request. Change the details to fit your situation.Comparison criteria sheet
Decision to make: [ ] Options under consideration: [ ] Must-have 1: [ ] Must-have 2: [ ] Preference 1: [ ] Weight (if useful): [ ] Preference 2: [ ] Weight (if useful): [ ] Disqualifying issue 1: [ ] Disqualifying issue 2: [ ] Option name: [ ] Meets must-haves: [yes / no / unknown] Disqualifier present: [yes / no / unknown] Score on preferences: [ ] Facts still to confirm: [ ] Criteria added after seeing options, and why: [ ] Final choice and who approved it: [ ]
Criteria that drift toward a favorite
Compare the criteria in the answer with your original list. Check for a must-have quietly weakened or an added criterion that changes the ranking without your agreement. A newly noticed criterion may be useful, but it should be proposed explicitly and applied consistently. Verify that disqualifiers are handled as you specified. Recalculate weighted scores if you used them. Keep missing information marked unknown rather than assuming it favors an option.
Ask another agent to check the result
Compare the recommendation with my original criteria. Flag added or changed criteria, weakened must-haves, ignored disqualifiers and inconsistent treatment of unknowns. Check the arithmetic for any weights I supplied. Separate a proposed new criterion from one already approved, without guessing why the assistant favored an option.
Test the recommendation
Ask which unresolved facts or reasonable changes in your preferences would alter the recommendation. Verify the important must-haves for leading options against their sources. If you revise a criterion after seeing the comparison, record why and compare all options again. Keep the final choice and any remaining uncertainty together.