CounterFact helps teams evaluate policy changes before live rollout.

Most teams already have decision logs. They know what action was taken, what context was available, and what outcome followed. CounterFact uses those logs to compare a proposed policy with the current one, then reports both the estimated change and the strength of the evidence behind it.

The product grew out of production machine learning and causal inference work on repeated decision systems: renewal offers, save treatments, routing choices, onboarding prompts, recommendation modules, and intervention policies.

We built CounterFact around one rule: a policy estimate should not stand alone. It should come with evidence checks, visible limits, and a practical next step.