AI is now predicting your next cancellation

10 Sept 2026 · 1 min

Watch on YouTube

0:46

In this monologue, Dan Novaes explains how companies employ AI-driven churn prediction to anticipate when a customer is likely to cancel a subscription. He outlines the behavioral signals that feed these models—login frequency, feature usage, support interactions, and sudden drops in activity—and describes how AI can flag at‑risk users before they act. The businesses can then automatically present tailored retention incentives such as discounts, free months, or cheaper plans, meaning two identical customers might receive different offers based solely on AI forecasts. Novaes raises the consumer‑side question of whether knowing that a firm can predict cancellation changes the bargaining power of a customer, suggesting that future negotiations may happen before a customer even decides to leave.

Chapters

Before you cancel a subscription, AI might already know you're leaving. Companies collect signals from how customers use their products— how often you log in, which features you stopped using, whether you've contacted support, or if your activity suddenly drops. AI can analyze those patterns and estimate which customers are most likely to cancel. That's called churn prediction. But here's where it gets interesting: a company doesn't have to wait for you to leave. If its AI thinks you're about to cancel, it could automatically offer you a discount, a free month, or a cheaper plan. 2 customers using the exact same service could eventually receive completely different offers based on what AI predicts they'll do next. For businesses, that's incredibly valuable. But for consumers, it creates an interesting question. If AI knows you're willing to leave, could threatening to cancel actually get you a better deal? Because in the future, companies may not wait for you to make a decision. They'll try to predict it before you do.