PostHog Autoresearch
Builds, evaluates, and deploys agent-driven predictive machine learning models in PostHog to forecast user churn, conversion, and feature adoption.
Builds, evaluates, and deploys agent-driven predictive machine learning models in PostHog to forecast user churn, conversion, and feature adoption.
PostHog Autoresearch empowers product and data teams to answer predictive questions like 'who will perform an action in the next N days' using automated, agent-driven machine learning. This skill enables Claude to configure model targets and populations, validate training datasets, initiate sandboxed agent training, monitor holdout and realized AUC metrics, and utilize resulting propensity scores directly in cohorts, feature flags, or SQL queries.