Local population context
Explore public statistics and appropriately licensed data to represent regional, age and lifestyle differences. Document coverage and assumptions.
AI POPULATION SIMULATION · RESEARCH
Taiwan Sandbox
AI consumer simulation research for Taiwan. We explore how different groups might respond to a product or service before ideas are tested in the field.
2026 AI 新銳選拔 · Initial review passed
Silicon Rising Co., Ltd. passed the initial review with its consumer sandbox proposal, publicly named Taiwan Sandbox. Announced August 31, 2026.
THE QUESTION
What could we learn before a field study?
New products and services begin with assumptions about people. We explore whether simulated populations can help teams compare those assumptions and focus their interviews, surveys and pilots.
Three research priorities
We aim to test whether local data and explicit population settings produce more consistent, testable comparisons than one-off model responses.
Comparable scenarios
Design agent-based experiments that vary a product concept or service scenario while keeping the comparison traceable.
Validation against observations
Plan comparisons with surveys or pilot observations. Examine subgroup errors, consistency and conditions where the approach fails.
Applications we aim to explore
Product concept research
Compare how simulated groups respond to a concept and identify questions for subsequent user research.
Service scenario comparison
Explore possible differences between offers or service journeys before committing to a pilot.
Research design support
Identify overlooked assumptions and groups, helping researchers refine sampling and evaluation plans.
Planned research outputs
- 01
Data and evaluation design
Document data sources, assumptions, baseline methods and an evaluation protocol.
- 02
Simulation prototype
Build a testable workflow for scenario setup, agent simulation and result comparison.
- 03
Pilot evaluation report
Compare simulated and observed results, report error patterns, and assess the next product step.
A research tool, with validation at its core.
Taiwan Sandbox is in research planning. The applications and outputs above are planned, not validated capabilities. Simulated responses should inform research hypotheses; they do not replace representative surveys or establish how real people will act.
Help shape a useful test.
We welcome market researchers, domain experts and businesses with a concrete research question, suitable data or a pilot setting.