Your churn dashboard might be accusing your product of something your payments stack actually did. When a renewal charge fails, customers can disappear even if they still want what you sell, and that “involuntary churn” quietly drains ARR, wrecks retention analysis, and forces teams to make the wrong fixes.

In this episode I sit down with Charles Rosenblatt, CEO of Butter, to unpack how subscription businesses can recover revenue from failed card payments using machine learning. We get specific about why many dunning programs rely on simple retry schedules, and how an ML-driven approach can choose the right moment for each invoice based on patterns in card and issuer behavior. Charles also shares how Butter works globally and plugs in as an abstraction layer on top of existing processors, so merchants can improve recovery without ripping out their current setup.

Then we look ahead: Butter is expanding from reactive recovery to predictive modeling that helps companies anticipate payment failures before they happen. Charles walks through real operational scenarios across telehealth, AI subscriptions, physical subscription boxes, and gyms, where a predictive score can guide whether to ship, whether to pause access, and how to reduce wasted CAC. We also talk about using anonymous card transaction variables without PII, and where AI is genuinely useful versus just hype in the payments industry.

If you run subscriptions, payments, billing, or growth, you’ll leave with a clearer way to separate product churn from payment friction and a roadmap for improving revenue retention. Subscribe to Leaders in Payments, share this episode with a teammate, and leave a review with your biggest churn question.