People close to the problem
Select respondents whose existing circumstances make the job, constraint, or decision relevant before the solution is described.
§ Use case · artificial respondents
Artificial respondents for product concept testing
The cheapest time to discover that a product idea is solving the wrong problem is while the idea is still words, sketches, and assumptions. Artificial respondents let the concept meet an audience before engineering hardens it.
- Ideas, feature concepts, and rough propositions
- Find objections before engineering
- Use real users when the concept earns the next step
Last updated
The decision
Research while the decision is still fluid
An early concept does not need a polished prototype to expose the assumption that everyone inside the team has started treating as true.
Describe the concept as clearly as you can today. Select respondents who live close enough to the intended problem to have meaningful reasons for accepting, rejecting, or reframing it.
Read the disagreement before you build. The goal is to learn which parts of the concept are obvious, which need evidence, which depend on a narrow audience, and which should be taken to real users next.
Why Antestrata is different
Respondents are synthesized before the research brief searches for them. The research selects people from the generated population rather than inventing personas to fit the task.
§ 02 · The panel
Build the panel around the problem, not the feature
A product concept becomes more useful to test when respondents are selected for the circumstances that create the need.
Use geography, work, household context, income, technology readiness, habits, attitudes, constraints, and other available characteristics to find people who can encounter the problem from different lives.
Antestrata selects those people from the existing population. The concept does not generate a set of convenient characters who already fit the team's hypothesis.
Sceptics and alternatives
Include people whose current workaround is good enough. They force the concept to compete with the real status quo rather than with an empty market.
Different adoption patterns
Compare early adopters with people who need proof, social validation, routine compatibility, or a clear practical advantage before changing behaviour.
Rare intersections
Use the generated population to explore combinations of geography, profession, language, household, or constraints that are expensive to recruit early.
§ 03 · Workflow
How concept testing works
Keep the concept honest. Test what exists, not the future product you hope the reader will imagine around it.
01
Write the concept
Bring the product note, feature description, rough proposition, sketch, or early flow with enough detail for someone outside the team to understand the intended job.
02
Choose people around the problem
Select respondents whose lives make the underlying need, trade-off, or existing workaround relevant.
03
Ask for reactions and reasons
Look for misunderstood jobs, existing alternatives, trust barriers, missing proof, irrelevant features, and reasons a person would do nothing.
04
Decide what earns the next step
Revise, narrow the audience, change the concept, or move the strongest unresolved question into real-user research.
§ 04 · Research questions
Concept questions worth asking early
The useful answer is usually the reason behind acceptance or rejection, not a score for how much someone likes the idea.
What problem do you think this solves?
Check whether respondents infer the same job the product team believes it is addressing.
What would you use instead?
Expose the real alternative: another product, a spreadsheet, a colleague, a habit, doing nothing, or accepting the problem.
What part feels unnecessary?
Find features that make internal sense but add complexity without changing the respondent's decision.
What would need to be true for this to matter?
Turn conditional interest into concrete requirements the team can validate before building deeper.
Planning your study
Artificial respondents cannot prove demand, retention, habit formation, or willingness to change behaviour in the real market.
Use them to improve the hypothesis before engineering and live research make that hypothesis expensive.
More use cases
Other work worth pressure-testing first
The product stays the same. The panel, research question, and thing being tested change with the business decision.
Pitch decks
Pressure-test a pitch deck with artificial respondents selected for investor-relevant backgrounds and decision patterns before real VCs and angels see it.
UX research
Use artificial respondents to pressure-test interfaces, onboarding flows, prototypes, and app journeys before recruiting real usability participants.
Landing pages
Test landing-page positioning, proof, trust, pricing context, and calls to action with artificial respondents before sending real traffic.
Pricing
Pressure-test price presentation, plan logic, value framing, premium tiers, and willingness-to-pay questions with artificial respondents before real buyers see them.
Onboarding
Use artificial respondents to find setup friction, unclear next steps, trust breaks, and activation problems in onboarding before real users encounter them.
§ Before the real users.
Let the idea meet resistance while it is still cheap
Bring the concept you are debating internally. Select the people, read the disagreement, and decide what deserves to be built or tested with real users.