How to Ask AI the Right Questions to Find Products That Fit Your Needs

Ask an AI tool for "a good laptop" and you will get a good laptop. It will be perfectly reasonable, widely praised, and quite possibly useless to you, because "good" is not information. It is a feeling.

The quality of what you get back depends almost entirely on the quality of what you put in, and most of us put in far too little. The fix is not complicated. It is just specific.

Replace adjectives with constraints

"Affordable" means nothing. ₹15,000 to ₹20,000 means something. "Lightweight" means nothing. "Under 1.4kg because I carry it on the metro daily" means something.

Every vague adjective in your request is a place where the system has to guess, and it will guess toward the average. The average is rarely you.

A quick rule: if your request contains a word that two people could reasonably interpret differently, replace it with a number or a situation.

Describe the situation, not just the product

This is the single biggest upgrade you can make. Products are chosen for contexts, and the context carries most of the useful information.

Compare "a good office bag" with "a bag for a daily metro commute, needs to hold a 14 inch laptop and a lunch box, gets rained on four months a year, and I would rather it did not look like a laptop bag." The second one has effectively already picked the bag.

Mention who it is for, where it will be used, how often, and what the environment is like. Indian weather alone rules out enormous amounts of what would otherwise look appealing.

Say what you do not want

Negative preferences narrow things faster than positive ones, and almost nobody offers them unprompted.

"Nothing with a strong fragrance." "No subscription required." "Not white, I will ruin it." "Nothing I have to hand wash." Each of these removes a large chunk of the catalogue instantly.

Dealbreakers are particularly worth stating up front, because they are the things that cause returns. If a product fails a dealbreaker, nothing else about it matters.

Be honest about your actual habits

This is where most people quietly sabotage themselves. We describe the person we intend to become rather than the person who will use the thing.

If you will do two skincare steps on a weeknight, say two, not five. If the treadmill is realistically going to see Sunday mornings, say Sunday mornings. If you want something low maintenance because you are not going to maintain it, say that plainly.

The tool is not judging you, and the honest version produces a product you will actually use. The aspirational version produces something that sits in a cupboard reminding you of a plan you abandoned.

Give a range and a reason

Instead of a hard ceiling, give a range and say what would justify the top of it. "₹3,000 to ₹5,000, happy to go to ₹5,000 if the battery life is genuinely better."

This is far more useful than a single number, because it tells the system what you value, not just what you will tolerate. It also lets it show you when spending a bit more actually changes what you get, and when it does not.

Tell it what you already tried

The most underused input of all. If you have bought something in this category before and it did not work, say so and say why.

"I tried a gel cleanser and it left my skin tight" is worth more than three sentences about what you want, because it rules out an entire family of products and tells the system something true about you rather than something aspirational.

The same goes for what did work. "I liked my last pair but the ear tips kept falling out" gives a precise target: same qualities, different fit.

Iterate on one thing at a time

If the results are not right, resist the urge to rewrite everything. Change one input and look again.

Usually a single answer was doing the damage. Maybe your budget was slightly too low for what you described, or a dealbreaker was cutting out the whole category you wanted. Changing one variable tells you which. Changing five tells you nothing.

Ask for comparisons, not just picks

A question that gets a much better answer than "what should I buy" is "what is the difference between these two, for someone like me?"

Comparison forces the reasoning into the open. Instead of a verdict, you get the tradeoff, and the tradeoff is the thing you actually need in order to decide. One has better battery, one is lighter. Now it is your call, which is where it should be.

Similarly, "what would make me regret this in six months" is one of the most useful things you can ask about any product. It surfaces the long term annoyances that reviews written in launch week never mention.

Three mistakes that waste everyone's time

Asking for the best. There is no best, there is only best for a set of constraints, and if you do not give the constraints you will get the most popular thing instead.

Asking too broadly. "Something for my mum's birthday" cannot be answered well. Her age, her interests, your budget and whether she already owns one can be.

Accepting the first answer. The first response to a vague question is a starting position. One round of refinement usually improves it more than the original question did.

A template you can reuse

I need [product] for [situation and how often]. My budget is [range]. The most important thing is [priority]. I definitely do not want [dealbreaker]. For context, [relevant habit or constraint].

Five lines. It works for a mattress, a pair of running shoes, a mixer grinder and a face serum. Keep it in your notes app and you will never write a vague request again.

When the tool asks the questions for you

Writing a good brief is a skill, and it is slightly unreasonable that shopping should require one. The better approach is for the tool to ask, because most people cannot describe what they want until they are shown two options and asked to pick.

That is how Nudge works. You pick a category and a product type, and instead of facing a blank box, you answer a few quick questions, each one framed as a simple choice. The questions change depending on what you are shopping for, so you are never asked about skin type when you are buying earbuds.