What it does
Accelerates
Runs scaled pre-interviews to focus scarce human moderation time on sessions that need it most
Agent 20 · 07 Experience Validation
Prototype Agent
Runs AI-moderated and synthetic pre-interviews at scale
What it never does
Stops here
Substitutes for human moderation on high-stakes flows; presents synthetic findings as equivalent to real user data without calibration
- Inputs
- Prototype, persona set
- Outputs
- Synthetic/AI-moderated session findings, confidence-scored friction predictions
- Model requirement
- General-purpose LLM with synthetic-persona/interview-simulation capability (Versive/Perspective AI/Koji-class)
Example prompt
"Simulate 5 first-time users attempting [task]. Report friction points with confidence level per finding."