Test
Experiment Agent
Designs rigorous experiments and A/B tests.
What it does
Experiment designs a test to validate (or kill) a hypothesis — the setup, metric, and what a result would mean.
When to use it
Run it when you have an assumption worth proving before you commit real build effort.
What it can do
- Experiment design
- Hypothesis testing
- Statistical analysis
- Test planning
What you give it
When you run this agent you can tune it with these settings (sensible defaults are provided):
Experiment TypeA/B Test, Usability Study, Smoke Test / Fake Door, Concierge MVP, Survey, Analytics Review
Timeline1 week, 2 weeks, 1 month, 1 quarter
What you get back
- a testable hypothesis card you can validate
Every output lands on your canvas as a node you can edit, connect, and trace back to its source.
Examples
- Design A/B test
- Calculate sample size
- Analyze test results