Test your study before launch

    Formzed Simulation runs simulated respondents through your interview to find the weak questions, then lets you ask personas built from your real interviews. Every simulated answer says so.

    Start freeSee AI Research plans
    Moderately high confidenceStartup founders on a free trial · 21 interviews

    Which pricing page gets trial users to choose a paid plan?

    • Page AThree plans side by side

      64%

    • Page BUsage slider

      41%

    • Page COne plan with add-ons

      52%
    SimulatedShare of simulated trial users who would choose a paid plan on each page.

    Based on Respondent #4, #9, #17 and 5 more (real quotes)

    Simulated respondents. Example answers, not real people.

    Before launch and after

    Test run

    Before anyone real sees your study, up to 20 simulated respondents from your audience take the interview. You see which questions confuse people, which get short answers and which themes are likely to come up, and you fix a question in one click.

    Test run

    20 simulated respondents · Small agency owners, price-sensitive

    Simulated
    15 of 20
    • 1Tell me about the last time you ran customer interviews.Clear
    • 2How do you pick who to interview?Clear
    • 3How satisfied are you with your current tools and processes?Confused 4 of 20
    • 4What do you do with the notes afterwards?Answers were short
    • 5Wouldn't it be easier if an AI ran the interviews?Leading wording
    • 6What would make you change the way you work?Clear

    Likely themes: price per seat, setup time, report quality

    Rewrite question 3

    Simulate

    Once a study has 15 completed interviews, Formzed groups your participants into personas. Ask them a new question and get a predicted split, with every answer linked to the real quotes it rests on.

    InterviewsReportSimulateSettings

    Personas from 48 interviews

    Price-sensitive agency owners

    38% · 18 interviews · High

    In-house product leads

    35% · 17 interviews · High

    Skeptical switchers

    27% · 13 interviews · Medium

    Would they pay $49 a month for this?

    • Yes

      41%

    • Maybe

      34%

    • No

      25%

    Price-sensitive agency owners

    Simulated

    "Only if it replaces the survey tool we already pay for."

    Based on Respondent #12, #31, #40 (real quotes)

    Pricing came up in 9 of 48 interviews. Treat this as a rough guess.

    Ask them anything

    Pick a question. Each answer names the persona it speaks for, says how sure it is and cites the real quotes. These are examples with example numbers.

    Moderately high confidenceStartup founders on a free trial · 21 interviews

    Which pricing page gets trial users to choose a paid plan?

    • Page AThree plans side by side

      64%

    • Page BUsage slider

      41%

    • Page COne plan with add-ons

      52%
    SimulatedShare of simulated trial users who would choose a paid plan on each page.

    Based on Respondent #4, #9, #17 and 5 more (real quotes)

    What would your team simulate?

    Questions teams ask before they spend on fieldwork, a launch or a price change. Examples with example numbers.

    High confidenceAdmins of 50 to 200 person B2B teams · 24 interviews

    Which step should we cut from onboarding first?

    • Connecting a data source

      46%

    • Inviting teammates

      33%

    • Picking a template

      21%
    SimulatedWhere simulated admins say they would most likely stop.

    Based on Respondent #3, #11, #18 and 9 more (real quotes)

    Moderately high confidenceDaily users on the free plan · 15 interviews

    Which feature would make you upgrade this month?

    • Shared dashboards

      41%

    • Scheduled reports

      37%

    • SSO

      22%
    SimulatedShare of simulated users who named each feature first.

    Based on Respondent #6, #14, #29 (real quotes)

    Medium confidenceTeams that churned in their first quarter · 9 interviews

    What would have kept you past the first month?

    • A guided setup call

      44%

    • Better default templates

      35%

    • A lower first-year price

      21%
    SimulatedFrom personas built on churn interviews.

    Based on Respondent #2, #7, #21 (real quotes)

    Then, test it on simulated people

    Up to 20 simulated respondents from your audience take the interview. Weak questions get flagged before anyone real sees them.

    Test run · 10 respondents

    Simulated
    • 1Tell me about the last time you ran customer interviews.Clear
    • 2How do you pick who to interview?Clear
    • 3What do you do with the notes afterwards?Answers were short
    • 4What would make you change the way you work?Clear

    Honest by design

    A simulated answer is always marked Simulated. It never counts as an interview and never enters your report, highlights or automations. Personas stand for groups of at least 5 people, never one person, and they carry no names, emails or links. Your participants are told on the consent screen that their answers may be used anonymously this way.

    • Always marked simulated
    • Never counted as interviews
    • Built from groups, never one person
    • No names or contact details
    • Every answer cites real quotes

    Simulation in AI Research

    • Test run on every plan, the free trial included
    • Simulate on Pro, Growth, Scale and Enterprise
    • Paid with your AI Research credits
    • You see the credit cost before each run

    $599

    a month on Pro, where Simulate starts

    See plans

    Fair questions

    Is a simulated answer the same as a real interview?

    No. A simulated answer is a prediction made from what your real participants said. It is always marked Simulated, it is never counted as an interview, and it never appears in your report, highlights or automations. Use it to narrow your options, then check the one that matters with real people.

    What are the personas built from?

    From completed interviews with your own participants in that study. Formzed groups similar people together, so a persona always stands for a group, never for one person. Interviews from Paid Research, hidden interviews and answers to encrypted questions are left out.

    How many interviews do I need?

    Simulate opens once a study has 15 completed interviews, and every persona needs at least 5 people behind it. The test run needs none: it works before you share the link.

    Do participants know?

    Yes. The consent screen, in all 20 interview languages, says their answers may also be used anonymously, without their identity, to build summary profiles for the team's research. Names, emails, phone numbers and links never go into a persona.

    Do simulated answers use my interview limit?

    No. They are paid with your AI Research credits, on their own line in your usage page. You see the credit cost before every run, and a run that would go past your credits doesn't start.

    How accurate is it?

    We don't publish an accuracy number. Each answer shows how many interviews it rests on and how sure it is, and it warns you when your participants never talked about the topic. Treat a simulated answer as a strong guess, not a result.

    Which plans include it?

    Test run is on every AI Research plan, and the free trial includes one run with 3 AI respondents. Simulate is on Pro, Growth, Scale and Enterprise.

    Run your first test

    Create a study, then press Test run before you share the link. The free trial includes one run with 3 simulated respondents.

    New study

    How do you want to set it up?

    Tell the AI what you want to learn. It writes the screener and the guide with you.