OpinionGenYou approve every change
The assistant proposes. Your brief, audience and questionnaire only change when you accept.
See the workspace
Rehearse your research.
Design a study in conversation, run it against a synthetic audience, and walk into real research knowing which questions matter.
How would UK adults who buy cleaning products react to an £8 refillable spray starter kit?
Synthetic samplen = 120 · seed 20417
Not very likely
“£8 up front feels steep for a spray. I’d want to know refills cost less than a new bottle.”
Generated explanation · not a human reason
P2 · Rating
How likely would you be to try the starter kit?
120 of 120 synthetic respondents answered · 0 omitted
Introducing OpinionGen
OpinionGen lets you rehearse research in conversation, test it on a synthetic audience, and bring only the questions that matter to real people.
Briefs usually get written the week fieldwork begins, and the first time anyone tests a question is after the budget is spent. OpinionGen moves that rehearsal to the start.
Studies / Refillable cleaning spray
Conversation
Brief
Assistant suggestion · review before accepting
- Objective
- Find which objections to a refillable spray are worth probing in human interviews.
- Concept
- £8 starter kit with a glass bottle and two refill capsules. Refills £3 each.
- Decision it informs
- Pick two or three objections and one message to take into interviews.
- Out of scope
- Sales forecasts and market sizing.
How it works
From a hunch
to a study worth running.
Five steps, each one reviewable. The assistant helps with the drafting. You make every decision.
- 01
Describe what you need to learn
Talk through the idea and the decision it informs. The assistant asks clarifying questions and proposes a brief. You accept, edit or dismiss it.
- 02
Define who you want to hear from
Set the population, sample size and attributes, or import your own respondent file. Each attribute is marked as your source or an assumption.
- 03
Draft the questions
Single and multi select, ratings, open text and simple branching. Ask for a draft, review each question, then save a version that can’t change underneath you.
- 04
Review before anything runs
See exactly what will run, with every input frozen, and set a maximum budget. Nothing launches until you confirm.
- 05
Read the results, then ask why
Distributions with honest denominators, filters and your own weights. Open any synthetic respondent to see their answers, or ask them a follow-up.
Built to be trusted
Synthetic, and says so.
OpinionGen is for exploration, so it’s explicit about what was generated, what was assumed and what it cost. Every simulated answer is labelled as exactly that.
Assumptions are labelled
Every audience attribute is marked as your supplied source or an assumed distribution.
Samples you can reproduce
Populations are drawn from a recorded seed, so the same inputs give the same sample.
Frozen inputs
Each run is tied to saved versions of exactly what you reviewed. Past results never shift.
Spend agreed up front
Every run has a maximum budget you set and confirm before launch.
Honest denominators
Results show who answered, who didn’t, and when a segment is too small to lean on.
A starting point for research, not a substitute for people.
Why OpinionGen
Better questions.
Before the budget.
BeforeWith OpinionGen
The brief
Written in a hurry the week fieldwork starts.
Shaped in conversation, reviewed and versioned.
The questions
Found to be confusing after the survey closes.
Rehearsed against a synthetic audience first.
Priorities
Every objection gets equal airtime in interviews.
A shortlist of objections worth probing with people.
Assumptions
Buried in a slide nobody revisits.
Labelled on every attribute and saved with every run.
Start here.
Then talk to people.
Shape a brief, define an audience and rehearse your questions before fieldwork begins.