Rehearse the study before recruitment.
Give a defined synthetic audience your draft instrument and source materials. Use the resulting discussion to decide what the human study needs to test.
Request Early AccessCatch instrument problems while revision is cheap
A participant session is an expensive place to discover that a question is ambiguous or that a stimulus omits the fact people need. AnthroSim lets a research team rehearse those materials with synthetic participants before it recruits people.
Set an audience, add the exact packet participants should be able to consult, and choose a discussion format. The moderator runs an orchestrated room and records the transcript, participation events, source use, and final ballots.
Example: climate-policy instrument review
A research team is preparing interviews about a proposed carbon-fee and dividend program. It creates 24 synthetic participants across supported U.S. audience groups, then adds the policy summary and draft interview guide as project sources.
The team runs four versions of the question order. In the saved records it can inspect where participants ask for missing cost assumptions, which terms they interpret differently, and whether a source excerpt supports a claim made in the room. Those observations become revisions and hypotheses for the human study.
What the current product supports
- One or more audience groups, with 1-50 generated profiles per group
- Orchestrated discussions lasting 5-120 simulated minutes
- Uploaded files, pasted text, or an imported page as bounded project sources
- Exact final ballots and a complete saved-run JSON export
What the research says
Our population study found that calibrated personas improved on naive repeated prompting across 74 ANES demographic cells. Several simpler prompting methods matched or beat the persona arm on those coarse, static questions. Our group discussion study also found that a mechanism which improved one model harmed another.
That evidence supports a narrow workflow: use AnthroSim to inspect a synthetic room, generate questions, and refine materials. Revalidate the engine after a model change, and use human participants for population estimates, causal claims, and decisions about affected communities.
Request Early Access
We're onboarding a first cohort: academic and applied researchers, market research firms, enterprise product teams, law firms, policy organizations, and teams planning sensitive or regulated studies.
Early access participants receive preferred rates and priority onboarding.