Experiments
Experiments Overview
Test product changes against persona-grounded agents before production rollout.
Oriyn experiments simulate how evidence-backed personas may respond to a proposed product change and return verdicts, convergence, persona breakdowns, and caveats.
When To Run An Experiment
- Before building a feature that changes user behavior.
- Before moving pricing, plan gates, signup steps, activation flows, or onboarding copy.
- When a PR changes a user-facing flow and an agent needs product signal.
- When stakeholders disagree and need evidence-backed tradeoffs.
What Happens During A Run
Hypothesis is normalized
Oriyn identifies the intended product change and the affected user behavior.
Relevant personas are selected
The system grounds agents in personas and evidence for the linked product.
Agents evaluate the change
Persona-grounded agents reason through likely reactions and risks.
Results are summarized
Oriyn returns a verdict, convergence, persona breakdown, and caveats.
Limits
Decision support, not live measurement
Experiments help decide what to build or test next. They do not replace analytics, A/B testing, usability research, or judgment when the stakes require real-world validation.