Experiment Decision Workspace

Computed diagnostics · retrieved policy · auditable decision

EvalRAG Evidence Desk

Should this experiment ship?

Ask a product experiment question, optionally attach a CSV, and inspect the computed diagnostics, retrieved playbook evidence, and policy-constrained recommendation.

Try a real decision pattern

Drop a CSV for computed SRM, lift, and segment checks
Retrieval scope · All playbooks

Compute

Validate experiment data and calculate SRM, lifts, uncertainty, and segment risk when a CSV is attached.

Retrieve

Search only the selected product experimentation playbooks and preserve the ranked evidence chunks.

Constrain

Apply explicit launch policies to confirm or override the draft recommendation, then record the full trace.