ChatGPT Prompts for Market Research That Work

The short version: ChatGPT is good at structuring market research, generating interview questions, drafting personas and spotting patterns in data you paste in, but it cannot browse live pricing pages reliably, cannot be trusted for exact market size numbers, and it will make up competitor details with total confidence if you let it. Use it to speed up the thinking and the writing, not to replace the digging.

What ChatGPT is useful for in market research

Market research has always had three slow parts: structuring your questions, making sense of messy data, and writing up what you found so someone else will read it. ChatGPT is strong at all three. It is weak at the part most people secretly want it to do, which is telling you facts about a market it has never been given.

So the prompts below are split into what it speeds up and what you should never ask it to invent.

Prompts for defining your market before you research it

Say you run a small accounting software company and you are thinking about launching a product for freelance photographers. Before you spend a penny on surveys, use ChatGPT to sharpen the question itself.

• "I'm considering a bookkeeping tool for UK freelance photographers who earn £20,000 to £60,000 a year. List the 8 assumptions I'm making about this market and turn each into a question I need to answer before I build anything."
• "Give me 5 ways this market could be smaller than I think, and 5 signs that would tell me I'm wrong."
• "Act as a sceptical investor and poke holes in the idea that freelance photographers need separate bookkeeping software from general freelancers."

This is the bit most market research articles skip. The first output from ChatGPT is never the useful one. The useful one comes from the second or third prompt, when you push back on its first answer and ask it what it got wrong or what it's assuming without evidence. Treat the first response as a draft, not an answer.

Prompts for building customer personas

Once you know roughly who you're researching, personas help you write better survey questions and interview scripts.

• "Build 3 customer personas for a freelance photographer bookkeeping tool: one just starting out, one established with 5+ years experience, one who's scaled to hire an assistant. Include their biggest money worry, the tool they currently use instead, and the one word they'd use to describe bookkeeping."
• "For each persona, write 3 objections they'd have to paying £15 a month for this tool."

Here's the uncomfortable bit nobody says out loud: these personas are fiction. ChatGPT is pattern-matching from training data and general stereotypes about freelancers, not from actual interviews with actual photographers. They are useful as a starting hypothesis and a writing aid, not as a substitute for talking to five real customers. If you skip the real conversations and build your product plan on AI-generated personas alone, you are guessing with extra steps.

Prompts for writing survey and interview questions

This is where ChatGPT earns its keep, because question design is a skill most people never learned and it is good at it.

• "Write 10 survey questions to understand how freelance photographers currently handle bookkeeping. Mix multiple choice and open text. Avoid leading questions."
• "Rewrite this question so it's not leading: 'Don't you wish bookkeeping was easier?'"
• "Write 8 open-ended interview questions for a 20-minute call with a freelance photographer about their bookkeeping habits. Order them from easiest to answer to most revealing."

Ask it to flag any question that's leading, double-barrelled, or likely to get a polite lie instead of a real answer. It's better at spotting bad question design than most people are, because it's seen thousands of examples of both.

Prompts for analysing data you already have

This is the single most useful application and the one most posts underplay. If you've run a survey through Google Forms or Typeform, paste the raw answers in and ask ChatGPT to find patterns.

• "Here are 40 open-text survey responses about bookkeeping frustrations [paste them]. Group them into themes, tell me how many responses fall into each theme, and quote the 3 most representative comments per theme."
• "What's the most surprising thing in this data that I might have missed?"
• "Turn these themes into a one-page summary I can share with my co-founder."

Step by step, this is how it works in practice: export your raw responses as plain text, paste them in batches of around 30 to 50 (much more and quality drops), ask for themes first, then ask for counts, then ask for quotes, then ask it to challenge its own grouping by checking for any response that doesn't fit cleanly anywhere. That last step matters because ChatGPT tends to force everything into tidy categories even when a response is an outlier.

Prompts for competitor research

This is where you need to be the most careful. ChatGPT does not have live access to competitor pricing pages, recent funding news or current feature lists unless you're using a version connected to the web, and even then it can get specifics wrong.

• "List the main categories of competitor for a freelance photographer bookkeeping tool: direct competitors, adjacent tools freelancers already use, and 'do nothing' as a competitor. For each category explain the real threat."
• "I'm going to paste in the pricing pages of 3 competitors. Compare their pricing tiers and tell me where the gaps are." [paste the actual text]

Notice the pattern: ask it to reason about categories and strategy, but feed it the actual facts yourself rather than asking it to recall them. If you ask "what does FreeAgent charge per month" without pasting anything, you might get an answer that was accurate two years ago, or close but wrong, stated with total confidence. For anything that needs current web information, a tool built for live lookup is the better first stop; our comparison of Perplexity and ChatGPT is worth reading before you decide which one to open for a given research task.

Prompts for sizing a market without faking precision

Everyone wants a clean TAM/SAM/SOM number. ChatGPT will happily give you one, built from estimates stacked on estimates, and it will look precise because it has decimal points in it. Treat any market size figure it gives you as a rough order of magnitude at best, never as a number to put in a pitch deck without checking it against a real source like a government statistics office, a trade body report, or Companies House filings for competitor revenue.

• "Walk me through a top-down and a bottom-up estimate for the UK freelance photographer market size. Show your working and flag every assumption."
• "What published sources would have this number, so I can check it myself?"

That second prompt is the one people skip and it's the one that matters most. Ask it where the real number lives before you trust the one it made up.

A worked example from start to finish

Say you run a small subscription box business selling specialty coffee and you're deciding whether to launch a version for offices rather than homes. Here's a realistic sequence:

• Prompt 1: "List my assumptions about an office coffee subscription business and turn each into a research question."
• Prompt 2: "Write 8 survey questions for office managers about how they currently buy coffee for the team."
• Prompt 3: Run the survey with 30 real office managers through a form tool, then paste the open-text answers back in and ask for themes and counts.
• Prompt 4: "Here are 4 competitor pricing pages [pasted]. Where's the pricing gap for a 10-15 person office?"
• Prompt 5: "Turn all of this into a one-page summary with the 3 strongest reasons to launch and the 3 strongest reasons not to."

Notice that ChatGPT never once told you a fact about the market on its own. It structured the thinking, wrote the questions, and summarised the data you fed it. That's the whole job, and it's a real time saver, often cutting the write-up stage from a full day to under an hour, but it only works if the input data is real.

Where ChatGPT fits alongside your other tools

If you're already using ChatGPT connected to your Google Drive or Slack through ChatGPT's app integrations, you can pull survey results or customer feedback documents straight in without the copy-paste step, which saves real time when you're running research on a regular cycle rather than once. And if you're writing up findings for a client or a pitch rather than just for yourself, the discipline in our piece on ChatGPT prompts for pitching clients applies just as well to a market research summary, because the goal is the same: get to the point fast and back every claim with something real.

If your research needs are heavier, academic-leaning, or tied to formal citation requirements, it's worth knowing that OpenAI has restricted access for some research-tier features, which we cover in our piece on ChatGPT for researchers. For most small business market research you won't need that tier at all, standard ChatGPT with good prompts and real data will do the job.

The mistake that wastes the most time

The single biggest time-waster I see in how people use ChatGPT for market research is skipping straight to "write me a market research report on X" without feeding it anything real. You get back something that reads well, sounds confident, and contains numbers and claims that may be outdated, invented, or simply generic. It costs you more time in the end because you have to fact-check all of it after the fact, or worse, you don't and you make a decision based on it.

Good market research with ChatGPT always has a human doing the actual information-gathering, talking to customers, pulling competitor data, running the survey, and ChatGPT doing the structuring, the pattern-spotting and the writing-up. The moment you flip that order, the research stops being research and starts being guesswork with a confident tone. If you want real hands-on support turning this into a repeatable process for your business rather than a one-off experiment, that's exactly the kind of thing an AI consultant for small business can help you set up.

Frequently asked questions

Can ChatGPT replace market research surveys?

No. ChatGPT can write your survey questions and analyse the answers you collect, but it cannot talk to your actual customers for you, and any "data" it generates about real people's opinions without a survey is invented, not researched.

Is ChatGPT accurate for market size numbers?

Treat any market size figure from ChatGPT as a rough estimate only. Ask it to show its working and name the published sources you'd need to check, such as government statistics or trade body reports, rather than quoting its number directly.

What's the best first ChatGPT prompt for market research?

Start by asking it to list the assumptions you're making about your market and turn each one into a question you need to answer. This exposes gaps before you spend money on surveys or interviews.

Should I use ChatGPT or Perplexity for competitor research?

Use Perplexity or another live-web tool when you need current facts like pricing or recent news, and use ChatGPT once you've gathered that information to analyse it, compare it and write it up. Our Perplexity vs ChatGPT comparison breaks down which to open for which job.