FAQ

Short, verified answers to the questions people ask most before and after their first study. If your question is not here, contact support.

How accurate are synthetic individuals?

On the public accuracy benchmark, iMario's synthetic audiences match real survey answer distributions to about 89% across 11 real populations and more than 900 questions from named public data. The honest bar is not 100%: rerun the same survey on a fresh sample of real people and the distributions agree only about 93% of the time, so 89% lands close to that human ceiling, and it rises above 90% on questions anchored to a recent survey. A single individual also answers consistently, agreeing with itself about 95% of the time across rephrasings, against a human test-retest ceiling near 81%. Every headline number is reproducible from the public data. See the accuracy benchmark and how to interpret confidence.

Can synthetic individuals replace real research?

No, and iMario is not designed to. Synthetic individuals give you a directional read in hours instead of weeks, which is ideal for exploring options, ranking variants, and surfacing objections before you commit. For the highest-stakes calls, precise magnitudes, or brand-new questions with no precedent, validate against a small real sample. Think of it as a way to decide which studies are worth running with real people, not a substitute for them. See how to interpret confidence for when to trust a read and when to validate.

What does a typical study cost?

A small study runs on the free credits every account starts with. Creating five synthetic individuals costs 100 credits (20 each), and interviewing all five on a guide of ten questions or fewer costs 150 credits (30 each), for 250 credits total. New accounts start with 500 free credits, so your first study is covered. See Pricing and credits for the full rate card and more worked examples.

How long does a study take?

Most runs finish in minutes. Interviews run in parallel across a pool of workers, so a small group completes quickly and larger groups take proportionally longer. Generating new synthetic individuals happens in the background, and you can leave the page and come back. Every run streams live progress while it works, so you always see where it is.

What languages does iMario support?

iMario works across languages. A Discussion Guide is written in the same language you use to describe it, and each synthetic individual's profile is written in their own native language. Research reports can be generated in Chinese, English, French, German, Italian, Spanish, Japanese, or Korean; Quick Test reports offer Chinese, English, French, German, Spanish, Japanese, and Korean. In Studio, audience quotes stay in the audience's original language so nothing is lost in translation.

Quick Test or Research Report: which do I run?

Run a Quick Test when you want fast, structured reactions to a piece of content like an ad, an email, or a social post. It scores each version across dimensions such as Appeal, Comprehension, and Action Intent, at 3 credits per respondent. Run a Research Report when you want a deeper interview, where your synthetic individuals answer a Discussion Guide question by question, at 30 credits per respondent for ten questions or fewer. See Quick Test reports and research reports.

How many individuals do I need?

It depends on your goal, not on statistics. Use one to five for depth interviews or rehearsal, ten to fifty for most studies, and fifty or more to cover a wide, varied audience. Adding more individuals buys broader coverage of the audience's diversity, not statistical significance. See How many individuals do I need? for sizing by goal.

Do individuals remember between sessions?

Yes. Each synthetic individual has a memory that persists across chats and interviews, so a conversation you had last week carries into the next one. Memories carry a strength label shown as Vivid, Clear, or Fading: vivid memories are recalled confidently, while fading ones come with hedges, the way real recall works. See Memory for how the memory graph works.

What data does a LinkedIn import use?

A LinkedIn import builds a synthetic individual from a public profile, using career history, posts, and communication style, at 20 credits per import. It reads the profile and recent activity, then generates a persona that thinks and responds in a similar voice. Only import people for legitimate purposes, such as yourself, colleagues who have agreed, or public figures you want to rehearse with. See Import from LinkedIn, Pricing and credits, and our privacy policy.

What is Studio, and how do I get access?

Studio is a fast way to check a piece of content with a synthetic audience before you publish it, returning named objections instead of a guess about performance. It is in beta and appears as Projects in the left sidebar once it is enabled for your account. During the beta, checks are free. See Studio overview to see how it works.

Can I share reports with people who don't have iMario?

Yes, on paid plans. The Share button creates a link viewable by anyone, and you can revoke it at any time. You can also export a report as PDF, Markdown, or PowerPoint, depending on your plan. Share links and export formats scale with your plan: Free can view reports in the app but cannot create share links or export, paid plans start at 3 active links with PDF and Markdown export, and Pro and Enterprise are unlimited. See Share and export and Pricing and credits.

Is there an API?

API access is available on the Pro and Enterprise plans. If you use Studio, the @imario tool lets you run a quick, directional reaction probe from your own tools through the Model Context Protocol; it surfaces likely objections rather than predicting an outcome (see how checks work). For what your plan includes, see Pricing and credits.

Next steps

• Support: reach the team when your question isn't answered here.
• Quickstart: run your first study in about 10 minutes.
• Pricing and credits: the full rate card, credit packs, and refund rules.