Put market trial and error into auditable virtual experiments
Users create a reproducible experiment project, configure synthetic audiences and situational stimuli, run simulations and obtain simulation prediction reports with interactive data, predictive indicators, uncertainty descriptions and version audit chains.
- 01Define simulation projectFill in the test object, target market, decision problem, and situational stimulus.
- 02Configure synthetic audiencesGenerate or maintain synthetic agent portraits, and record data sources and versions.
- 03Run a virtual experimentPresent situational stimuli to a synthetic audience and record reactions and decision-making events.
- 04Review forecast reportsGenerate simulation forecast reports and review risks, uncertainties and decision recommendations by designated personnel.
Simulation forecast report
Results, sources, failures, approvals, and rollbacks are saved by project.
No business results yet
Please fill in the target market, portrait requirements and data source description first.
Data use boundary
Please do not upload or fill in real personally identifiable information, unauthorized social data, customer lists, mobile phone numbers, ID numbers, precise addresses, or original records that can identify individuals. Data source descriptions are only used to audit project assumptions and compliance boundaries; the system will mark the report as a simulated forecast rather than a real user survey. Current 10 Role Review Supplement: Data source descriptions must be minimal, structured, and document only source type, authorization/disclosure/desensitization status, and a brief summary; no individual-level raw data may be collected. ; Sensitive fields should be minimized in history, approval, export, logs and model calls, and should be desensitized or summarized. ; When deleting a project, the processing of project input, synthetic portraits, simulation runs, interactive events, prediction reports, export files, approval records, and rollback snapshots must be clearly covered; if security audit metadata is retained, the field range, period, and desensitization method must be stated and must not be used to reconstruct project content. ; Actions with usesWeb=false by default are not allowed to be retrieved online, and are not allowed to call social platforms, advertising platforms, CRM, BI, production, sales or customer database business APIs. ;OpenAI Responses API/video76-api should be expressed as a model processing infrastructure call and shall not be expressed as a business external system integration or real market data source.
Data retention
Project input, portrait version, simulation run, interaction event and report version are retained for 180 days by default to support audit recurrence; users can delete the project in the workspace, and the site will no longer display relevant reports and workflow status after deletion.
Human responsibility and rollback
The output of this product is a simulation prediction based on synthetic agents and user input scenarios, and is only used as a reference for market research, advertising testing and business discussions. The results may deviate from the real market, real consumers or actual sales performance and should not be construed as true survey conclusions, deterministic sales forecasts or automated business decisions.
Rollback only means restoring the project status, simulation run selection, report version or approval status within this site to the previous version; since the product is not connected to external production systems, rollback will not undo any advertising, production scheduling, procurement, CRM, BI or other external system operations.