Prompts for market research

What it is good at

• Structures a competitor or audience analysis fast once you supply the raw data.
• Surfaces positioning gaps that are hard to see from inside your own brand.
• Turns messy customer feedback or reviews into a clear list of themes.
• Generates a sharp list of research questions before you start, not after.

Where it falls short

• Without live browsing enabled, it cannot see current competitor prices, reviews or news, supply that data yourself.
• It can present a plausible-sounding but unverified claim as fact, treat every output as a hypothesis to check.
• It has no access to your actual sales or customer data unless you paste it in.

How to use it for creative work

1. 01

   Supply the raw data, don't ask it to recall it

   Paste in competitor websites' text, real reviews, or your own sales figures, don't rely on the assistant's memory of a market, which may be outdated or wrong.

2. 02

   Ask for a framework before an answer

   Request the analysis structure (e.g. positioning map, SWOT, jobs-to-be-done) before asking it to fill that structure in, so it doesn't wander into unstructured opinion.

3. 03

   Frame outputs as hypotheses to test

   Ask explicitly for 'hypotheses to validate' rather than 'the answer', which keeps you from treating a plausible guess as verified fact.

4. 04

   Use it to find the gap, not confirm what you already think

   Ask specifically what's missing from the market or what competitors aren't saying, rather than asking it to validate your existing idea.

5. 05

   Verify anything time-sensitive independently

   Prices, reviews, and news dated within the last few months should always be checked on the live source before you act on them.

Ready-to-use prompts

I'll paste text from three competitor websites: [paste]. Build a comparison table across: positioning statement, target audience implied, pricing model (if stated), and one clear differentiator each. Flag anything unclear rather than guessing.

Here's what these competitors offer: [paste summaries]. Based only on this information, what audience need or complaint does none of them seem to address? Give this as a hypothesis to validate, not a confirmed gap.

Here are [number] customer reviews for [our product/a competitor's]: [paste]. Group them into 4-6 recurring themes, note whether each is mostly positive or negative, and quote one representative line per theme.

Audience: [description]. Their main frustration: [frustration]. Competitors' typical claim: [summarise]. Draft 3 positioning statement options that differentiate from that typical claim, each one sentence.

I want to learn [research goal] from our customers. Draft 6 survey questions: 2 multiple choice, 2 rating scale, 2 open-ended, ordered so easy questions come first and sensitive ones come last.

Here are open-text survey responses: [paste]. Summarise the top 3 recurring points, note roughly how many responses mention each, and flag any single unusual response worth investigating separately.

Build a SWOT analysis for [our brand/product] using only this information: [paste: recent sales notes, customer feedback, competitor info]. Do not add assumptions beyond what's stated; mark anything you're inferring rather than reading directly.

We're considering pricing [product] at [price]. Here's what we know about competitor pricing: [paste, if known]. And here's what we know about our audience's stated budget: [info]. What questions should we answer before committing to this price, rather than telling us if it's right?

I'm running discovery interviews with [type of customer] about [problem area]. Draft 8 open-ended questions that avoid leading them toward our existing product, focused on understanding their current behaviour and frustrations.

I want to research whether [trend] is genuinely growing in [market/industry] or just visible in my own feed. List the specific questions and data sources (reports, search data, sales data) I should check myself, rather than answering the trend question directly.

Questions people ask

Can AI tell me current competitor prices?

Only if you paste them in yourself or the tool has live browsing enabled and you confirm it checked a current page, otherwise treat any pricing claim as outdated or invented.

Is this a replacement for real customer interviews?

No, it's best used to structure the questions and analyse data you've already gathered, not as a substitute for actually talking to real customers.

How do I avoid confirmation bias in the output?

Ask explicitly for the gap or the counter-argument to your idea, not validation of it, framing outputs as hypotheses to test rather than answers keeps this in check.