Skip to main content
Tell Humanize what you want to learn and it builds everything for you — discussion guide, screener, and settings.

​
How the AI Study Builder works

1

Describe your research

Tell Humanize what you want to learn in plain language, and pick the study type and language. You can also attach materials, such as an existing discussion guide. Humanize reads PDF, Word, Excel, PowerPoint, OpenDocument, RTF, CSV, TSV, text, Markdown, and images.
2

Choose your participants

Let Humanize find participants from our panels, or bring your own. Then set how many people you need and who they are. More on recruiting.
3

Set up your screener

Humanize suggests screener questions based on your goals, plus the welcome screen participants see first. Add, remove, or tweak them. More on screeners.
4

Review the discussion guide

Humanize writes the questions for each topic. Edit anything that doesn’t feel right. For a focus group, session length (15 to 90 minutes) and group size sit at the top of the guide. An interview shows its estimated length instead. More on discussion guides.
5

Launch

Review everything and click Launch to submit your study. Then start recruiting from the Recruitment tab.

​
Study types

Most studies are focus groups — several participants in one live session, reacting to each other. Two other formats are available: Choose the type when you create the study. It stays fixed after that.

​
Study language

Choose your participants’ language when you create a study. You can pick from these:
  • English: US, UK, Australia, Canada, and India
  • French: Canada and France
  • Spanish: Spain and Mexico
  • Portuguese: Brazil and Portugal
  • German, Italian, Dutch, Japanese, Russian, Hindi, and Thai
You write the guide and read the results in your workspace’s language. Participants see the study in theirs. Each language a study offers gets its own translated guide and screener, and each panel recruits in one of those languages. A launch in a language that is still being translated waits until the translation is ready. Each language also has its own moderator voice. Open the voice button beside a language to pick one, and press Preview to hear it. Every language starts on its default voice. A session speaks in the voice of the language it runs in. Results read in your workspace’s language too. When a session ran in another language, its transcript, quotes, and clip captions show translated by default, marked as translated where it matters. A toggle beside each one switches to what the participant actually said. Regional variants of your language, such as Australian English in a US English workspace, are never translated. The AI moderator’s voice supports considerably more than that. If you need a study in another language, get in touch — the moderator can speak any of the following, and we’ll tell you what adding it involves: Arabic, Bengali, Bulgarian, Chinese, Croatian, Czech, Danish, Dutch, English, Finnish, French, Georgian, German, Greek, Gujarati, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Kannada, Korean, Malay, Malayalam, Marathi, Norwegian, Polish, Portuguese, Punjabi, Romanian, Russian, Slovak, Spanish, Swedish, Tagalog, Tamil, Telugu, Thai, Turkish, Ukrainian, and Vietnamese.

​
Who can see your study

A study starts with your workspace’s default visibility. To change it when you create the study, open the + menu in the composer and choose Workspace visibility. Public means everyone in the workspace. Private means only people you add, and Admins. Private suits sensitive research or work you’re not ready to share internally. Workspace Admins can open every study, private ones included. The study’s creator and workspace Admins can change who can see it, and add or remove people at any time. Everyone else who can open the study sees who has access but can’t change it. Anyone added to a private study can remove themselves. Switching a private study back to public asks you to confirm, since that exposes it to the whole workspace.

​
Tips

  • Start with the AI builder. You can always edit the generated guide afterward. It’s faster than building from scratch.
  • Be specific about what you want to learn. “How do people feel about our onboarding?” gives better results than “tell me about users.”
  • Don’t over-scope. One clear research question per study. If you have three, run three studies.

​
Drafts and publishing

Edits save as a draft first — a dot on the builder header shows when you have unpublished changes. Save Draft keeps your draft. The first time, Launch submits your study, and you start recruiting from the Recruitment tab. After that, Publish Changes makes your edits live. You can rework a guide without affecting a study that’s mid-recruitment. Every published version is kept. Open the version menu in the builder to see the history, preview any earlier version, and restore it if the rewrite went the wrong way.

​
Study lifecycle

Once you launch, your study moves through four stages:
Finalize your screener before you launch. Once participants start answering it, changing it means different people were asked different questions — which breaks segments, quotas, and any comparison across your sample.