15 prompts to steal for your next study
Our research assistant sits inside every study. It has read every interview transcript, the published report, the codebook and the screener data, and it answers your questions in plain language, using your own data. Every number comes back as "X of Y" with a citation you can click straight into the interview behind it, and when the data is silent it tells you so instead of inventing an answer.
Here are 15 prompts worth stealing, from the quick orientation questions to the ones that do real analytical work, including a set that reaches across every study you have ever run. Copy them, change the topic to yours, and see what comes back.
Get oriented
1) "What is this study about?"
Scope, sample and topic in one line, before you open a single interview.
Find the evidence
2) "Find me quotes from people who don't trust these products, in their own words."
Pulls authentic verbatims on any theme, each with a citation that opens the exact interview.
3) "What share of participants already use the product, and who are they?"
Turns a vague sense into a number shown as "X of Y", with a chip that opens the matching interviews.
4) "What are the biggest pain points?"
Surfaces the strongest points of friction across the whole sample, with the weight behind each one.
Slice and compare
5) "Compare adoption barriers between the 18-22s and the 26-29s."
A cross-tab of two segments on the same question, so you see where they split.
6) "Go deeper on this finding: why aren't more people adopting?"
Takes one finding and expands it into a full answer built from the report, the codebook and the verbatims.
7) "Set up a coding dimension that groups people by their main reason for buying, then code every interview."
Re-slices the whole study around a new angle on demand, with no manual tagging pass.
Pressure-test it
8) "Cost is the biggest barrier, right?"
State a belief and it checks the data, correcting a wrong premise instead of agreeing with you.
9) "How does our brand stack up against its main competitors?"
Gives the real competitive set from the study, with the criteria where you win and where you lose.
Across all your studies
10) "Across all our studies, what do we already know about price sensitivity?"
Pulls one theme together from your whole back catalogue, so you build on what you own instead of starting from scratch.
11) "Which of our past studies touched on sustainability, and what did each conclude?"
Finds the relevant studies by topic and gives you each one's conclusion, resolving the study names itself.
12) "What did the packaging repositioning study conclude about premium formats?"
Recall a specific past study by name, no ID needed, and get its conclusion on a precise question.
Turn it into output
13) "What are the top three barriers, back each with a quote, and what should we do about the biggest one?"
Bundles several asks into one message and returns a single answer that handles all of them.
14) "What are the top recommendations from this study?"
Frames the strategic implications as "the data suggests", always with the counter-evidence attached.
15) "Make me a slide from this finding."
Turns a finding, or the whole report, into PowerPoint, ready to drop into the client's template.
The pattern
Every one of these works the same way: ask in plain language, get an answer you can click into, down to the interview behind the number. Take the prompts that fit your next study, change the topic to yours, and let the data answer.
Frequently asked questions
What can the Verso researcher agent do?
It answers questions about a study in plain language, using that study's own transcripts, report, codebook and screener data. It can quantify, pull verbatims, run cross-tabs, test a hypothesis, benchmark competitors, recall findings across past studies, and generate slides, with a citation behind every number.
Can it answer questions across several studies at once?
Yes. When more than one study is in scope, it resolves the study names itself and scopes each answer correctly. You can ask what you already know about a topic across your whole back catalogue, or pull the conclusion of a specific past study by name.
Does it ever make up answers or quotes?
No. It answers only from the data in scope and says so when the data is silent, for example "no participant mentioned that". It does not invent quotes, counts or competitors, and every figure links back to the source interview.
Can it build slides from a study?
Yes. Ask it to make a slide from a finding, or to turn the whole report into a deck, and it generates PowerPoint in the client's template. You can also ask which findings would make the strongest slides before generating anything.
What kinds of prompts work best?
Plain-language questions work best, the way you would brief a research analyst. You can bundle several asks into one message, state a belief for it to check, or point it at a single finding to go deeper. Swap the topic in these 15 prompts for your own.