You have 1,200 EU businesses in an early-access pool, heavy-tailed chat volumes, and expected 20–30% initial subscription take-up. Design a launch test to estimate impact on (1) subscription revenue per business and (2) Resolved-Within-24h Rate. Choose and justify between a cluster RCT (business-level or region-level), a stepped-wedge rollout, or a quasi-experiment. Detail: unit of randomization, stratification variables (country, size, baseline volume), sample size/power and Minimum Detectable Effect under overdispersion, interference risks (shared customers across businesses), compliance and adoption handling (ITT vs TOT with IV), sequential monitoring/alpha spending, and a pre-registered primary analysis. Provide the full analysis plan (estimands, modeling choices, covariate adjustment, heterogeneity by region/vertical, missing data rules), and a decision rule to roll out, iterate, or stop.